Slashdot Mirror
Are Expensive RDBM Systems Worth The Money?
Death of Rats asks: "What are the actual advantages of high-end Relational Database Management Systems (such as Oracle) over the free or low-priced solutions? Is it worth paying the 10 to 30 thousand dollars (or more) for Oracle, when you can do (as far as I know) everything it can by using ASP with Access/MSSQL for Windows, PHP/MySQL for Linux, JSP with JDBC connections for most platforms, or any number of other free or cheap systems? Is there a considerable speed increase, or is there enough more functionality to warrant the outrageous costs? Are some of the technologies more suited to different tasks? Any time I have needed to use databases, the open-source solutions have done fine, so I am skeptical."
There is no one-solution-fits-all answer to this question, as it really depends on the requirements of the task at hand -- which will in turn determine what one needs from an RDBMS. I'm sure there are tasks out there which better suit the high-octane RDBMS systems, but what tasks do readers feel still need that kind of power? Has the distance between Enterprise RDBMS servers and the free alternatives shortened at all?
Sort of. Its based on berkeley db. Its not exactly fast in the concurrent case (which is the only case that matters for a database server, as opposed to an embedded database library like BDB).
Interesting that you appear to be fairly well informed, but seem to think that DB2 is not a "competitor" in the enterprise market. DB2 hasn't done particularly well in the mid- and low- end Enterprise market that SQL Server and Oracle do well in, but when a company gets large enough, the inevitably go to DB2 because it can run on big iron. If you've got a million employees, you are not using Oracle on Solaris to maintain your HR database. Oracle is the best Database there is in the market it serves (as much as I hate to give Asshole Ellison any credit), but there is a whole level at the high-end that Oracle just isn't capable of handling - that's where you see DB2 keeping the books.
Yes, but would you rather run Oracle at a flight control terminal, or mySQL?
I've been an Oracle DBA and developer for almost 14 years. Yes, it's hideously expensive, but if you're doing mission-critical stuff, there's nothing better. Having said that, why would anyone spend so much $$$ on Oracle when the free alternatives exist? I've been playing around with MySQL recently, and there are two key features that aren't there: -Transactions -Row-level locking I've been using Oracle for long enough that I simply cannot imagine having to write an app without these two things available. Yes, I know people do it every day, but having these two things makes your life soooo much easier, and your system soooo much more reliable (e.g., no corruption of data, greater concurrency, etc.) From my standpoint as an Oracle user, going to MySQL is like revisiting the mid-1980s. Back to the future, anyone? Oracle had row-level locking in 1988, and has "always" had transactions (at least since version 5, and probably before). The idea of an open-source database is very exciting, and I like what I've seen in MySQL so far. I hope we can get to a point that we can have features that users of commercial products take for granted. Yes, most of Oracle's loudmouthed advertising has to do with scalability up to Sun E10K levels, 24x7, etc. - I'm not talking about those things. I'm more interested in the things I have gotten used to over a period of 10+ years and just can't live without. (Bitch/gripe/moan - if I were a better coder, I'd contribute to the open-source community. But I suck, so I won't.)
Most of the gigs that I've landed on have over-paid for a database technology. An overwhelming number of people just buy Oracle because it is Oracle. I see very very few people using the real-value add features of Oracle which leads me to think they could get buy with a cheaper, leaner competitor (Sybase, Informix). Oracle Licensing is VERY, VERY complex -- if you are compliant, you are paying alot. Not all applications need a feature rich database. If you dont need transactions or triggers and you can architect the application to cache result sets in memory in the application server tier -- why not use something cheap like MySQL or Sybase?? Having implemented some rather substantial web sites (b2b, retail commerce), I can assure you that a small dot-com style business can get by quite well by saving money and choosing a less expensive database technology than Oracle.
Reliability, scalability, availability, support, throughput for complex applications, security, concurrency, availability of knowledgable programmers and administrators, availability of programming tools, availability of data warehouse/decision support tools, and about a thousand extra features that will *never* appear in open source DBMSs. (Some people might say "don't use the product-specific features", but in my experience, that's a political position that cuts no ice in real-world situations.) In short: if you have to ask this question, you shouldn't be making decisions about DBMS selection.
The old phrase was "No one ever got fired for buying IBM. This Microsoft thing is new, new, new...
MS was nowhere on the map of officer-level folks until MS Office 4.2, which, I think, was late 1994.
We all have short memories, trained by rapid devel/deploy cycles, and an industry with 14 month lifecycles. The IBM quote was common in the 80's, and generally was used in reference to folks like DEC, DataGeneral, Amdahl, etc...
Jeremiah
"Flyin' in just a sweet place,
Never been known to fail..."
Huh? With open source, YOU can fix it yourself, or hire any number of companies or persons that are familiar with the code to fix it. It could be done in a few hours (assuming a simple problem) or maybe a few weeks for a complex problem.
With closed source, you're STUCK with the DB vendor. Granted, folks like Oracle BETTER be responsive to urgent needs, but what if they're not? Then you're completely screwed.
Having said that, i agree that Oracle is probably a better solution if you NEED the highest of the high end. But for 95% of projects, I'd feel quite comfortable implementing PostgreSQL. It has a great feature set and a knowledgeable set of developers who seem committed to supporting it commercially as well as in their free time.
I am pretty certain that as of version 11.9 Sybase supports row level locking.
However, one always had the option of setting the page size to limit the number of rows that might be included in a page lock that Sybase had used.
When do you need a high-powered RDBMS (like Oracle) over PostgreSQL or MySQL? When you need high-throughput, high-availability access to your data. Just as Linux won't scale to a 64-way machine, MySQL is hardly the solution for a multi-node redundant load-balanced database.
For *most* needs, PostgreSQL is a perfect solution. For anything that is enterprise-essential (or other buzzword-compliance), the Oracle RDBMS is a better solution.
Plus, Oracle Developer is a great RAD environment.
Microsoft is to software what Budweiser is to beer.
that's because banks have been using IBM since the dawn of time, and banks aren't exactly ones to move quickly when it comes to what stores their 'money'.
Find a bank, PERIOD, in the US, that doesn't use AS/400's somewhere along the line. (even those that outsource their storage and transactions to places like Alltel do this)
MySQL is small, fast, lightweight, suprisingly stable, is open source, and is extremely well supported.
Even people who have commercial Oracle support contracts claim that MySQL is better supported.
If you don't appreciate the extremely responsive mysql support mailing list, my company will gladly enter dedicated support contracts in NYC.
You sue the local Oracle support/dealer. If you have critical database and you NEED to get it up again (The Database that is) and your local Oracle support guy/dealer/whatever doesn't do his/hers job, well, I think he/she would be in a pile of shit! You would need an armageddon type event to ever haul Larry's ass in court, if that's what you mean ;)
It all comes down to the acid test
summarised in this url: http://www.shingray.com/town/database.html
ACID stands for Atomicity, Consistency, Isolation, Durability.... It means that a transaction on your database (a select, insert, update or delete) is done and stays done without any side-effects on other running transactions. As soon as you get in an area where you want to sell things( think E-commerce..:) ) you want a transaction to be complete or to be rolled back... Oracle or DB2 (and I believe Postrgresql) offer such functionality to extremes... mysql doesnt. period...
Okay, let's say there are 8,000,000 people in North Carolina. And let's say you have a 3,000,000,000,000 byte database. That amounts to 375KB worth of data for every man, woman, and child in the state.
What exactly do you keep in this database? Compromising pictures of every citizen?
-- Brian
The most rabid believers in American Exceptionalism are the exact same people whose policies are destroying it.
I'm pretty sure something like this must have been mentioned by someone before, but still:
How feasible is it to build a database engine with SQL support into an OS kernel? Are there any brave souls somewhere attempting this? A lot of the features of a multi-tasking OS can be used without having to redo the same thing again at the application level. Some things like the scheduler would need to be seriously augmented maybe?
With your average database servers running pretty much just the database itself hogging all the resources the OS is kind of just there to provide hardware abstraction isn't it.
Kill'em! Kill'em all!
First off, some terminology. An RDBMS (relational database management system) like Oracle or MSSQL server or DB2 is a substantially different beast than Access or MySQL which are simply RDBs (relational databases). MySQL offers some minor facets of a full-blown database management system but comes NOWHERE close in full-featured transaction systems with full integrity constraints and in-system procedures. PostgreSQL has a larger subset but lacks clustering and extreme scalability.
Now, for the average website, a full-blown RDBMS is overkill in the extreme. That much I'll agree with. (And Oracle in particular has some REALLY annoying features in their JDBC driver like only being able to deal with 4000 characters at a time in an insert [without resorting to tricks and hacks reaching around the standard JDBC layer] - whereas MySQL can handle much larger chunks of data.)
So, for websites that are web-centric and not dealing with internal systems, MySQL or PostgreSQL are probably fine.
James
- Intel Only? Umm I have it on a Sparc Right now.
You have Windows 2000 with MS-SQL 2000 running on a sparc? Or do you have some back-reved version of NT and MS-SQL that will never be updated?Thought so...
To be fair, some simple queries did have lower latency on MySQL, but when the same queries were run in parralel (a few hundred clients), MySQL quickly fell behind Oracle. Further, if your website involves a lot of multi-stage updates to the database, you had better have Oracle. With MySQL, your options are: slow and contentious, or fast and corrupted.
The point is: in terms of performance, MySQL will start out faster than Oracle, but Oracle scales better. Running the command-line SQL program is not a valid benchmark. Also, don't trust the benchmarks on MySQL's website, they are not representative, or verifiable. While Oracle may take X seconds to run an operation with one user, it is reasonable to assume it will X seconds with a thousand users. The same is not true of MySQL.
In terms of pure features, MySQL is pitiful next to the might of Oracle. Oracle supports features like transactions, distributed transactions, replication, standby databases, PL/SQL, Java, and SQL92 support. MySQL supports: transactions, sorta. Plus, Oracle's backup solutions are incredible. It's nice to be able to have the server rewind the whole database to yesterday morning.
As others have mentioned, the support is phenomenal. I had a problem one time, so I called up Oracle, and got one of the techs on the phone. Eventually, they determined that I had run into an undocumented, platform specific bug. So, I sent them a number of trace files, and they had a programmer look them over. Within a few hours, I had a work-around. I compare this to my experience working with open-source programmers. If they don't think the problem is a big deal, they won't fix it. With Oracle, you really get your money's worth on technical support.
Yes, I'm still a junky. Are you still a bitch?
'nuff said.
I honestly CAN'T believe the poster mentioned Microsoft Access.
I read the story about what happened at Ebay. Apparently their sysadmin(s) are absolutely incompetent. Their database got corrupted due to a known Solaris bug for which a fix had been available for 6 months (!) prior to the incident. Also, they did not have redundant servers for their production database or a cluster of any kind. Ironically, they were just weeks (or days?) away from putting a second DB box in production for load balancing and failover...
___
___
If you think big enough, you'll never have to do it.
. . .even worse, clearly the /. moderators haven't either. Otherwise, they wouldn't even bother wasting our time with such a stupid question. Oh wait, "rememeber children, there's not such thing as stupid answers, only stupid people"
** Martin
... the isp i worked for had a full oracle contract and after bench marks, setup, implementation, et all, decided to go with mysql. The major advantage to oracle is that if you have alot of cash, they can provide support 24x7x356 support... then again for the kinda cash it requires for that you can hire a 'good database admin' and pay them and still save a few dollars. *shrug* depends how much of an inhouse shop you are i guess.
Shadus
I would argue that products such as Oracle require as much or more babysitting. I have developed applications that use Oracle for 5 years and would not recommend it unless you are willing to pay for DBA services. Oracle is a complex product that will not perform without 1+ competent DBAs. Is the same true for MySQL? PGSQL? Interbase? BTW-One of the 'selling points of Interbase is that:
Most SQL database server products require expensive MIS staffs to install, tune, and manage them. The InterBase design doesn't require hours of maintenance or a PhD in InterBase tuning. When your applications must run without constant supervision or when your desktop database runs out of steam, InterBase is the clear choice.
--
"In the land of the brave and the free, we defend our freedom with the GNU GPL."
"You're gonna need a bigger boat." - Chief Brody
Interesting thought. So how much money is it worth when I find one of the numerous bugs I Oracle, even though Oracle never (rarely) fixes any of them? So I have to spend several days coding around these problems? What good is the 'service' that I am paying for?
--
"In the land of the brave and the free, we defend our freedom with the GNU GPL."
"You're gonna need a bigger boat." - Chief Brody
Considering how slowly Slashdot is loading today, I'd say yes.
gimme a break, no competition
Well -- it used to be "nobody got fired for buying IBM" (back in the 70s and 80s).
Since IBM hasn't been such a mover and shaker in 10 years, it's more appropriate to Microsoft.
These kids today -- they think time began in 1991 with Window 3.1 and DOS 5. *sigh* :)
Potato chips are a by-yourself food.
This is a very good point. I thought it sorta went without saying, tho.
I've had some success with Oracle on Redhat (and, actually, works very well for a development or test environment), but for real-live deployment, you go Sun hardware with Solaris, and you pay for the hardware support as well.
(I'm no huge fan of Sun hardware, but these day's it's the best of a bad lot -- shame on SGI for having the bad grace to slowly spiral down the drain. IRIX sucks, but SGI hardware is nice. Of course, this is IMHO.)
Potato chips are a by-yourself food.
Don't get me wrong -- Oracle isn't perfect. It's just among the best of the lot. And, Oracle DBAs (well, good ones) are gods among men, which tends to make Oracle look better than it really is.
For example, Oracle's OCI extensions have some of the most brain-dead things I've ever seen. Why do I care about Oracle's desire to use pointers in >2K CLOB entries? Screw that, put stuff whereever you damn well feel like it -- that's what I paid $80K for the license for! I don't wanna have to think about how Oracle stores stuff. That's it's job, not mine.
But overall, I trust Oracle more than PostgreSQL. However, for my latest projects, my clients won't be able to pay for the Oracle license, so I'm using PostgreSQL -- and not really all that terribly scared about it, I'm just more comfortable with Oracle.
Potato chips are a by-yourself food.
Well, I'm really tounge-in-cheek, since I'm a believer enough in IBM to be a shareholder. I'm not sure of the exact date, but IBM removed the head from their ass and started making excellent stuff.
I've got enough IT friends who get all misty-eyed over IBM mainframes to know better than to dis them for real.
But, IBM isn't the omnipresent god-like figure they used to be, back when they had a corporate song-book and all the reps wore blue suits and ties.
Potato chips are a by-yourself food.
My conclusion: always prototype with a small sql database: if you use a database at all, you don't want to limit yourself to three users (:-)). Be prepared to move to a big one if the rollout is large enough to warrent it.
davecb@spamcop.net
Since I have started working with interbase I've found it is a lot less complicated than oracle, MySQL, ms sql, or pgsql. Although not as powerfull as oracle, I'd put it up agains ms sql any day, and prefer it to mysql.
But it really does depend on the number of users and the load that is going to be on the server to really make a good decision. While a small company, up to about 50 users I'd probably go with interbase. But more than 50 I would want oracle and possibly middle ware servers.
Overall interbase has been very easy to get going.
If ignorance is bliss, the world is full of blissful people
Yeah, this is true, I pretty much leveled with the tech support guy, told them I had taken over the system from someone else and honestly wasn't exactly sure what was going on. Gave them all the info they needed and they helped me learn and fix the problems. A lot of it made sense when I was finally sent to Sybase Adaptive Server Enterprise training. After that I could call tech support tell them the problem, and throw all the stuff they would ask for up front. Amazingly it has helped in getting the problem resolved quicker. As far as being hung out to dry, the boss couldn't do that very easily as I was taking on a ton of responsibility and managing to make the systems quite a bit more reliable.
Although there have been a few times where I got a first level engineer who was lucky to answer the phone, he couldn't set up his voice mail correctly, and didn't know how to read the case notes. Luckly for me I called back way after hours and got another first level engineer from Australia. He pointed me in the right direction and got the problem resolved. Luckly these have been rare at companies such as Sun, HP, and Sybase. Although it seems much more common at Dell. It's like they have an unbreakable script they cannot deviate from. It took me a month to get them to ship me a new ethernet card, even after shipping it back to Dell 3 times for the same problem. I would imagine Dell has some great support engineers, although I haven't had the pleasure of working with them.
As far as the job thing goes, hopefully in the next few days I'll have an offer in writing from Sun Microsystems, as I know UNIX far better than I do databases.
"If you insist on using Windoze you're on your own."
Well a few years ago I was thrust into being a DBA and an HP/UX admin. At that time I didn't know the first thing about databases, nor much about admining HP/UX. I did know Solaris and how to admin Linux fairly well at that point. Needless to say I was one of those "clueless nitwits" who called support quite a bit. Over time I learned from the helpful and intelligent support staff members. Honestly if it wasn't for them I couldn't have gotten my ass out of the fire so to speak. Since the systems that I inherited were poorly managed by the previous admin, partially due to his lack of training, and partially to having far too many responsibilities.
While I agree with Ian Wolf that "...a competent Oracle DBA who stumbled across a hung database would immediately check for a log switch, missing file, or lack of free space. For DBA's this knowledge should be as second nature as breathing." I think that a large portion of this problem rests with the companies who tell someone with a bit of computer experience, but definatley not a competitent DBA to run their Sybase, Oracle, DB2, or whatever RDBMS system. In the case of some companies they simply won't pay for a real DBA (in the case of my current job). And in many of the same cases, the company won't send their people to training... So what's left? Pretty much that leaves the tech support 800 number.
"If you insist on using Windoze you're on your own."
I'd agree with the idea that IBM was still a mover and a shaker if the context was still in mainframe computing, but have you ever tried websphere?
It is, by far, the worst application server I've ever worked on, including ColdFusion. (At least ColdFusion did what it said it does without breaking). That takes them out of the mover and shaker league in my book, as a solid application server is not that difficult to code up, but is very necessary for middle ware going forward.
DB2 could use something as easy as PL/SQL as well, but that's a minor point compared to the loathing that I've developed for WebSphere over the last few months of building a massive e-commerce site on it.
That last point gets it back on topic as well, if we're talking about high-end databases, I'd put my money behind Oracle given lots of past experience - I think other posts sum it up well. I don't think DB2 is worth the same amount of money though, and IBM disagrees - so I'd vote with my wallet and send the bucks to Oracle.
There really is no substitute for stored procedures, tight/fine-grained transaction control and database-generated sequences.
Yeah, the most annoying part of the standard mail template is that all your email is sorted oldest->newest and you have to scroll down to get the new stuff. Quickly fixed with a simple edit of the template but that should be a menu choice for all the non-"programmers" (can we really count editing forms as programming? hehe). Ah well... Notes is a biotch to program. Just a major pain in the rump compared to anything else...
Yeah, I did a focus group on Lotus Notes and kinda ended up dissing IBM pretty bad... Felt bad about it afterwords when I remembered all the open source stuff they are contributing and the fact that they play fair. Ah well, it wasn't the end of the world, I just said they were too conservative I suppose (living in their own towers and all that)... I don't that is true these days anymore either.
Postgres 7.x is supposed to be dramatically faster... I haven't had a chance to do any benchmarks - has anyone else? Personally I'm hoping Postgres takes off and we can switch to it from MySQL (for triggers, transactions, etc).
There is a big advantage to Oracle, even in the face of others big commercial databases: writers do not block readers. Or is it the reverse? :-)
Anyway, nobody blocks nobody, which allows for a large number of long-running queries to run in parallel.
Interestingly enough, PostgreSQL has that feature, while no other commercial database has it.
--
As many have already mentioned, being able to
deal w/ TeraBytes of data is something Oracle
does well.
Also, there are many other facets:
- Parallel Server
- Standby database
- B*Tree indexes
- reverse key indexes
- bitmap indexes
- cost based optimization
- Rock solid transaction processing
- unequaled recoverability
- ability to recover up to last commited
transaction
- point in time recoverability
Oracle is damned expensive, but they are by far
the market leader, and the technological leader
when it comes to RDBMSes.
If you have a small database as a back end for
a website, don't waste your time and money.
If you need the most reliable, robust, tunable,
scalable, and recoverable database money can buy,
Oracle is the answer.
There is no equal.
-Mark
P.S. Yes, I'm an Oracle DBA. No, I don't work
for them.
There are 10 types of people in the world; those who understand binary and those who don't.
I agree that anyone questioning this issue should read these books.
Also, understand that databases come in sizes and scales, too. Indeed, this often comes up in solicitations for job applicants to work with such databases. One such value point is whether or not table unloading/reloading is a "reasonable operation" or not. There are databases of tens of gigabyte size, typically up to 100 million rows (assuming some average bytes per row), for which doing maintenance can involve an unload or, at worst, some kind of fast export. Then there are bigger ones for which this is an entirely unreasonable thing to do for which maintenance must be done using the same SQL interface that updates and queries on it are done.
One thing the Greenspun books can show you is that there are other ways of using databases in the construction of Web sites besides as simple containers for accounting, purchasing, or clickstream data. In particular, one can serve pieces of Web pages from the database and images stored as blobs. This is especially easy if one uses a facility like MetaHTML. This does, of course, have the side effect of placing those pieces fetched that way beyond the reach of search engines. OTOH, some sites want to achieve precisely that effect.
Jan Theodore Galkowski, (Oo) http://www.smalltalkidiom.net/ MySQL,PHP,ETL,SQL,MinGW C, and plucking the Web
It goes even one step further. Oracle + what platform? I don't know any successful enterprise system that DOESN'T use Solaris.
Twelve-and-three-quarter inches. Unyielding. This wand belonged to Bellatrix Lestrange.
But this discussion has been on /. before... I just can't seem to find it in the Database :)
This discussion is on /. every 3 months
That may be true... Too bad the licensing agreements for all major closed-source database products absolutely prohibit the publishing of any benchmarks... Good luck trying to prove it!
-Hunter
RateVegas.com - Vegas Reviews
Ever try explaining to a manager that your web app is running on "MySQL?"
Even the die-hard Microsoft bigots at my place of employ will say "ok" to an app running on Linux. Linux is hip. "MySQL" is not hip. Just the name makes managers ask, "it'll be in Oracle when it goes production, right?"
I know this is insane, but names are important and MySQL doesn't have the right one. Not when managers are involved.
I don't think that was the case here. I think the person asking the question just does not know much about databases. Anyone who tries to compare Oracle with MySQL needs a heap 'o learnin.
Open-source databases, in general, haven't endured the real-world conditions that Oracle, DB2 and MS-SQL have.
Interbase has. It's paid its dues, been around the block, been sold for hundred of thousands of dollars. And now it's free. See the firebird project on sourceforge for the open source version. Interbase has been used by Motorola for 911, and the army for tanks. It is mission critical.
Compatibility: Interbase / firebird had its SQL interface built to the SQL spec. It's probably impossible to be completely in line with the spec (too much open for interpretation), but it's a very good effort.
Interbase has been around for a long time, supports SQL-92, stored procs, triggers, generators, and has a super-spiffy multigenerational architecture that allows readers and writers to work without blocking each other. Yes, really. Check out Firebird, get involved, make it what you want in an RDBMS.
MSSQL is probably the most useable enterprise database but is much slower than Oracle or DB2.
What we tend to do is prototype in MSSQL and then switch to Oracle (for solaris) and BD2 (for AIX) and sometimes MSSQL on win2000.
It's much easier to develop for MSSQL as get very good debuging for stored proceedures (similar in style to VC++), and we have to rollout on multiple databases (so we have to mess about porting databases anyways.)
DB2 is a pig to setup, but is reliable and scaliable. Oracle is complex, but probably the best RDBM.
(Oh, PS Anyone know how to get Oracle to export a SQL script that to create the current database (like you can when you generate a script in MQSQL ) Thanks)
Not to mention ego-less-ness, oustanding visuals, and increased apprecation to music!
Your Momma's so fat she makes emacs look like nano!
uh... yes.
repeat after me.
"Oracle is to RDBM's as IBM was to mainframes."
Now do you understand the point? Thank you.
-josh
actually, if anyone remembers, the key to this whole challenge was a very subtle detail (and this is very important evidence to the contrary for the person that stated "if they'd open it up to informix, sybase, etc, they'd have lost".
oracle was the ***only*** RDBMS at the time that supported materialized views (the ability to take a physical snapshot of a particular query's result set and have the optimizer *automatically* decide when to use it to speed up the retrieval of the result set without a change in the query). this little detail made it such that NOT A ONE of the other database vendors could even come close to outperforming oracle's solution. oracle just performed a temporal shift of the i/o and computing horsepower necessary to solve the query.
it was simply rigged from the very beginning.
would you expect any less from larry? he is the #2 man in hell (wishing he were #1) after all.
Peter
Databases support transactions. MySQL doesn't support transactions. MySQL isn't a database.
"But wait!" you say. "I know of databases that don't support transactions!" That is because you are completely wrong. You may know of programs that don't support transactions. Maybe they store stuff. And maybe they let you query stuff. But they are not databases. I like to think of them as 'toys.'
Oracle is a database. BerkeleyDB is a database. MySQL isn't.
As for the "user error" response - it does happen I'm afraid but let me say this. I get so many calls from people who are supposedly IT professionals who have no clue at all that it's real hard to stay un-cynical - but I try. But I also make sure that I get them to check the simple things first. Sometimes people hear that as patronizing or me blaming them but it's not - it's a troubleshooting procedure. I know for a fact that for every guy who's offended by me asking if any of the filesystems are 100% - there's two who haven't checked it. (And 5 others who just say no and we move on.) Believe me we'll quickly get to the non-trivial problems if that's what it is.
My gripe is the exact opposite. I answer with exact honest and precision with all the basic questions they bother with. Where I get pissed is the *days* of "analysis" they go through and the fingerpointing they get into. If the first-line guy doesn't have an option when I've answered his obvious questions, the ticket ought to be escalated to someone with more knowledge and better problem solving skills, and not kept with the drone who regurgitates the fingerpointing excuses fed to them by their line managers who have call-resolution numbers to make.
Are you the same idiot that I replied to last time? You sound even dumber. The reason why is, in order to get that $1 million, you'll be paying a lot more than that in software and Oracle consulting fees. That's why knowledgeable people call it a scam.
Cheers,
What are you talking about? The top entry was a cluster of twenty-four 8-cpu machines. Two that topped Oracle's highest score was a cluster of twelve 8-cpu machines. (Of course, the setup running Oracle cost $9.6 million and the two SQL Server setups cost $4.2 million.) What's this "100s of 'small' Xeon machines" crap? Go lie somewhere else, please.
You talking about Larry's $1,000,000 scam pretty much says it all about your knowledge of enterprise databases.
Cheers,
For being at a "severe disadvantage," Microsoft doesn't seem to be doing too badly. SQL Server 2000 holds the number 1, 3, and 4 positions on the TPC-C performance list (number 2 is DB2 7.1, but like the rest of the top 4, it's a Win2K system). The first place entry more than doubles Oracle 8i's 5th place entry, and gets $19.80/tpmC compared to Oracle's $43.30/tmpC. SQL Server 7 and 2000 also hold all of the top 10 price/performance positions (all that were listed).
Cheers,
Oh come on! Anyone who calls Oracle a "piece of [garbage]" is not to be trusted. Now I'm willing to admit MS-SQL may be wonderful, but I'm not going to believe it when it comes from someone who calls oracle crap in the same breath.
I'm also willing to admit Oracle is hard to use. But in terms of performance, Oracle is king. Usability is not normally the main criteria when chosing an enterprise-level system.
If this is all you have to say, I'll just assume you like MS-SQL because it's easy and you like windows in general because it's easy. Dat don't make it good. Show me the benchmarks!
-=Julian=-
"producing a db that actually implements the ACID properties reliably is very, very hard"
Actually, it's not that hard. What is hard, is producing a database engine that implements ACID properly *and* performs well for large datasets and large user counts. Naive implementations of ACID tend to result in performance bottlenecks when you try to scale up.
I tried going to Oracle's web site and typing "guarantee" into their search box. I got back:
Database Log In Failed
TNS is unable to connect to destination. Invalid TNS address supplied or destination is not listening. This error can also occur because of underlying network transport problems.
Verify that the TNS name in the connectstring entry of the DAD for this URL is valid and the database listener is running.
Kinda puts the whole reliability thing in perspective...
I have first hand experience of a large .com migrating their warehouse management system from MySQL to Oracle just 'because'. It cost them millions of dollars in software, staff education and Oracle consultants and they took a big performance hit after it was rolled out. The code also became more difficult to debug because what was just pure perl was now alot of PL/SQL too. One could argue that MySQL doesn't scale as well as Oracle, but hey - slashdot get's upwards of a million page views a day, and it seems to be coping just fine.
Well if you're comparing MSSQL to Oracle, then no. MSSQL is much cheaper, and is almost, if not as fast as Oracle. Access isn't even a competitor in this market. Access, like FoxPro, etc are desktop db apps and not really meant for high-availability web apps.
As for open source solutions I don't have enough experience to honestly compare. I've been using MSSQL / ASP at work for years and they really rock.
I use PostgreSQL / Perl on my personal home Linux website and it's pretty quick and efficient.
Not everyone on the other end of the phone is a drone.
I concur with this.
Commercial RDBMS should be robust, configurable and scalable. These large scale databases provide options and capabilities beyond what is available as open source as far as I know, such as data partitioning, blobs, roll back recovery, tuning, large volume sizes.
From a personal work view, I want to deliver apps without having to debug and fix problems with my toolset. I acknowledge that there will always be bugs and fixes aren't always timely. However if there are problems these are handled by the software company providing the tools.
My view is horses for courses. If the app is only storing information for simple to moderately complex retrieval and manipulations then open source is possible since its at that level. For heavy duty database work, commercial RDBMs are currently the way to go.
Let the flaming begin.
/* Andrew Fong - rogue programmer */
Then there's the case of the company who's DB is worth $1M/hour when it's down. Who are you going to sue if MySql goes down for a day? Disclaimers aside, Oracle actually has some money you can go after.
Who are you going to sue if Oracle goes down? Does anyone have information about the odds of winning lawsuits against Oracle/Sun/Microsoft? Yeah, they have money to win, but that means they have money to defend. It's almost certain that somebody using these products did something that was not recommended. Which, of course means that any problem would not be the fault of the vendor.
Smart folks use a product for what it can do for them, not for what it can do for their lawyers.
Information is not Knowledge
Well, I've blown my cover in another post so I might as well mention it here where it's also relevant: the application I was involved with was Oracle Designer and related tools. It's a pretty complicated beastie, but is capable of supporting many concurrent users allowing them all to share model data and not lose any of it.
--
Dunx
Dunx
Converting caffeine into code since 1982
In all the time I was developing applications on Oracle I never saw the database die. It never lost data through anything but bugs in the client application, and we only found one real bug in the server code in all that time (which was not data threatening, and was fixed before production release of the database as far as I recall).
Even if the application had not been an internal Oracle development, I don't see how it could have been developed on another company's database engine simply because Oracle's in-database features were so powerful. It would have been impossible to ensure data integrity across multiple user sessions doing the kind of complex transactions which this application required if there had not been available triggers, stored procedures, and lots of other things which I do not understand but knew were necessary.
I've done some client application development using other databases, and it's never been as secure an experience. Access (not really comparible, but my only other major database client development stint) lost data all the time, and MS SQL Server (albeit an older version) lacked many data integrity features which were essential.
Right at the moment, if you're building big systems (either due to data size or sheer complexity) Oracle is way to go.
Oh, and hello Gerard.
--
Dunx
Dunx
Converting caffeine into code since 1982
I certainly would not claim that Oracle is a perfect company - as I say, I am a former employee, and there were a number of reasons why I chose to leave. I shall not burden the readership with all of these, but certainly misgivings about technological direction played a (small) part.
I just don't see Oracle ever using embrace/extend that's all, and (FUDdish sales tactics aside) there is a tendency to compete more on performance and features than Microsoft will do.
--
Dunx
Dunx
Converting caffeine into code since 1982
I'm curious as to your reasoning here - Oracle is undoubtedly an aggressive company, but any ambitions of world domination are much more narrowly focussed than those of Microsoft.
More relevant to the topic under discussion, many people have mentioned MS SQL Server as a cheaper alternative to Oracle RDBMS; Oracle port their database to any number of platforms, including Linux. Microsoft will never, ever do that.
--
Dunx
Dunx
Converting caffeine into code since 1982
Since I no longer work for the company I obviously have no influence on fixing any bugs, but I might be able to help. Might be best off line, though, if you want to mail me.
--
Dunx
Dunx
Converting caffeine into code since 1982
Umm, I can't see that any dbms can guarantee recovery of your data if your hard drive dies nor is that the meaning of "Durable". (It might if you use a particular style of logging to multiple drives but that's another story).
Durable means that if you yank the cord on the machine 1ms after the commit returns then the transaction is safe.
The only thing that the various open source RDBMS systems do not do is scale nicely to supprt large - really large databases - like say 200,000,00 - 300,000,000 tupules (entries for those not versed with DB theory). When the open source DB's like Postgres/MySQL and such do large data-sets and ave the same data recovery features like Oracle, then wwe'll be looking at taking over the world. until then sure a small business with 20,000 entries will run fine on Postgres on MySQL. But finacials for a large company - no.
Have fun with yer Data Bases
Why do they need to be in your city? This is the internet. They can connect remotely. Plus are you a big corporation? Pay them the $50K and see.
Reference: Intro To Database Systems, 8 ed., C.J.Date, Chapter 14.2 "Transactions" and 14.3 "Transaction Recovery".
Not to troll, but this just proves you don't know how to use it. I've developed in Access since 2.0 for small environments (~25 concurrent users) and have never, ever lost one bit of data. Granted, I'm not saying Access can be compared to Oracle in terms of scalability, but for single office environments, it is more than adequate for the task. Certainly better than the upper limit of one you impose.
Examples:
Now, do I need one RDBMS that does all that? No. I'm sure one of several DBs could be hacked into doing one of the above specific jobs just fine. As others have said, if you only have one job to do, pick the best tool for the job and hopefully it won't be Oracle (for the cost).
If, as any large IT shop with a diversity of database scenarios will tell you, it's hard to beat Oracle. One DBA, with a fixed background, can administrate the whole deal. Hardware outdated? As long as you don't need a recent version of Oracle, you can get it to work.
BTW, if any Oracle geeks know how to do a COMPLETELY distributed disconnected database, let me know. I have a scenario with a mandatory air gap - no "periodic refresh" as with the travelling salesman. I can sneaker-net ASCII files, but that's it. Thanks.
Yes, RI comes at a price, but only when you actually use it. Whether you can afford the cost of RI depends on your particular application.
Not having true transaction support does make MySQL fast for _simple_ web applications.
All said and done though, it's not just the simple queries that people do. Frequently customers end up doing complicated stuff. Building a compiler and optimizer for producing a good access plan for arbitrary queries isn't really just a matter of coding. There has been two decades of active computer science research that's gone into it !! On top of that, you have to build the infrastructure into your code for being able to have the right features. And this is not even saying anything about fancy features like federated technology.
Disclaimer: my salary is paid by a company that vends a commercial RDBMS - I'm obviously biased.
3. Data integrity Oracle's got more locks, rollbacks, and ability to survive power failures and such, than MySQL etc... by MySQL's own admission IIRC.
We used to use Postresql 6.3 on our site and switched to Oracle to please our PHB. With Postgresql we had:
1) Rock Solid stability
2) Instantaneous response to moderately complex queries
3) simple administration
4) zero cost (with the exception of a really bright dba)
With Oracle we have
1) Rock solid stability
2) Instantaneous response to moderately complex queries
3) complex administration
4) $55,000 up front plus $20,000/yr for support plus a really bright dba.
The support, by the way, is via repeated incorrect responses from their web site.
Other comments:
1) Postgresql is a better comparison to Oracle than is MySQL. MySQL is a great product for what it does but is not (yet) a full-featured RDBMS.
2) Access is a datafile on a PC. You can rw it with a library built into an application.
3) MS SQL server is a great mid-upper tier database if you want a MicroSoft solution.
4) Oracle has incredible scalability. Very few organizations will ever need that scalability. Look at your organization and decide where you can best get a ROI on your money.
>Can mySQL do that? My experience with support on >open-source solutions has been good, but not good >enough to bet my business on getting a problem >fixed in a few minutes instead of a few hours or >days...
:)
Hahaha! Getting through to oracle support takes more than a few minutes. Let along talking to someone with a clue, having them open a ticket, then get back to you days later saying it is a but/unimplemented feature.
Needless to say I'm not really a huge fan of oracle these days.
At my last job, I worked with both MySQL and Oracle extensively.
There's a difference. A real difference.
At first you'll think they're the same, but as you start getting more advanced with your queries and commands you want to run, it starts becoming obvious. I'd have to do big work arounds and spend hours doing things that would be a few calls in Oracle.
Eventually we moved the MySQL machien to PostgreSQL and that helped a bit. But I still preferred oracle.
Not to say that I dislike MySQL or PostgreSQL. They're good applications. But Oracle one of the few expensive things that is really worth the money.
I really doubt oracle is slower than mySQL for stuff like this. I just think that it's like using a "too good tool" for a simple job. In oracle you can really fine tune your queries and your server itself to speed them up alot. The issue is that for a small website, it might not warrent the costs needed for oracle.
actually, it seems to work pretty okay in my experience... whats the problem?
Juln
Some things to think about: replication, stored procedures, triggers, subqueries. You may not always need all of these features, but often they can be very helpful. For example, imagine you have a web based application that (unfortunately) uses a couple different languages for it's front end. Some processes, like adding a new user to a table of users might require that you cross check data in two different tables. If your database supports it, you could write a stored procedure that performs that crosscheck for you. If it succeeds, it allows the insert statement. If it fails, the database returns an appropriate error. This way, you don't have to perform that logic in your program, and even better, you don't have to code it in two different languages. This is a simple example. I'm sure others will provide many more reasons. In short, there are still pretty significant differences between a database like oracle or db2 and something like mysql. Don't get me wrong, mysql definitely has it's place, but like so many things, there is no single database that's appropriate in all situations.
Actually, while giving a "case-study" of Veritas Cluster Server at the Vertex 2001 conference, an Ebay admin said they were running two fully loaded Sun E10ks running Oracle on the backed, and were looking at splitting the Oracle database into multiple chunks. No postgres, etc. =-)
In all of this it's worth remembering that large companies spend a lot of money on PR to make not only you but your boss comfortable with the product, and a lot of that feelgood PR has nothing particular to do with the facts.
I always wondered why Oracle Forms is and remains such a dog, and why they never integrated DEC/RDB into Oracle proper when they bought it ages ago. Currently they sell them both as seperate products (Oracle and Oracle RDB).
In my experience (a bit dated on Oracle now) Oracle was alright, but wasn't anything special, the row level locking was pretty ordinary and was contained in a lot of other products at least 5yrs before Oracle signed on to it (for a large price addon as I recall).
And when many say 20x7x365, I must admit I've never seen one that really worked. They all seem to have a convenient memory lapse, and forget the 3days downtime where they screamed at people to get it back on the air while upgrading the software, that sort of thing. Mostly it seems to be done by duplication of the site and some sort of SQL patchup/merge after the problems are all over.
The free databases will get better, and I think eventually replace the costly commercial ones. Everyone will still make a bucket out of support etc, since after people loose their data once they usually start bothering to do backups (either online or otherwise)..
If it worries Larry that his income will fade a little, he'll just have to console himself that the same is happenning to Bill, his rival in the most pointlessly richest man in the world competition.
Just my 0.52c worth.
Cheers
Mark
Too bad the TPC doesn't measure uptime. Speed and benchmarks are worthless if the system doesn't stay up.
Dave
The second step towards release certification is running an ever-growing suite of regression tests. Some verify functionality, some verify scalability, recovery capabilities, etc. Not only do these tests have to be run, the results have to be analyzed, any discrepencies noted, and determination made whether the result changes are a problem with the new database or with the test suite itself.
Finally the beta partners get their hands on the RDBMS software, and start beating on it with applications whose requirements might stress aspects of the RDBMS that aren't covered by the in-house regression suites. Again, those issues have to be categorized and addressed, either before release or as a known issue that will be patched.
If it sounds like a lot of work, you're underestimating the scope of the efforts involved. It is a mind-boggling effort whose costs often far exceed that of the actual software development.
That is the most important thing you are paying for when you buy a commercial RDBMS.
Beyond that, you need to consider what the vendor (open source or otherwise) designed the RDBMS for. Some focus on data collection or OLTP performance (Sybase, for example), some on simple online queries (most do well at this), some on complex MIS reporting (DB/2 and Oracle), and a very few on data aggregation and warehousing (often not even an RDBMS.) If the vendor can't tell you what their design goals are, but tries to convince you it can do anything, you had best be prepared to either spend a lot of money tuning your application for the RDBMS, or on hardware in a potentially futile effort to address performance issues.
Bottom line is to use what suits your needs. To paraphrase an old saying/questionaire:
You can have it:
Pick any two.
I do not fail; I succeed at finding out what does not work.
What are the actual advantages of cars over the free or low-priced solutions? Is it worth paying the 10 to 30 thousand dollars (or more) for a car, when you can do (as far as I know) everything it can by using a big wheel, a skateboard, a bicycle, or any number of free or cheap public transportation options? Is there a considerable speed increase, or is there enough more functionality to warrant the outrageous costs? Are some of the technologies more suited to different tasks? Any time I have needed to use a car, my Razor scooter has worked just fine, so I am skeptical.
What are the actual advantages of girls over the free or low-priced solutions? Is it worth getting a girlfriend, when you can do (as far as I know) everything it can by using Playboy, "American Pie," alt.binaries.*, or a cantaloupe? Is there a considerable speed increase, or is there enough more functionality to warrant the outrageous costs? Are some of the technologies more suited to different tasks? Any time I have needed a girl, my hand has worked just fine, so I am skeptical.
Well either, since they'd only be running the drinks machine.
perl -e 'fork||print for split//,"hahahaha"'
perl -e 'fork||print for split//,"hahahaha"'
This is a classic case of picking the right tool for the right job.
If all you want is some structured storage for a low usage intranet app, and don't need transactional capability, then something like mySQL kicks butt. It's quick, reliable and not resource intensive. We use Bugzilla as a tracking tool for our Engineering department at maybe 500 hits a day and it rocks.
If you have a ton of mission critical data and you need serious availability, and you need transactional atomicity, when forget the marketing crap and TPC numbers, you have *only* two choices: Oracle and DB2. For each, you have two core platforms (Oracle on Solaris or HP/UX, DB2 on AIX or OS/390). For a serious big ass DB, there is NO substitute for proven technology on big iron. Period.
I am the world's biggest open source advocate, and we use both FreeBSD and Linux extensively in our production setup, but for the DB we use gold old Oracle on Solaris. It's not fancy. It's not cool. It's not cheap (but not violently expensive). But get this - it WORKS.
I know all the advocacy weenies who've never run a real production service will come out and say how Oracle 9i supports clustering on Linux, or (depending on flavour of weenie) how SQL Server can get a bazillion TPC's if you cluster 4,756 NT boxes together. It's all irrelevant horseshit. When your ass is on the line for uptime and data integrity, you'll be glad you bought a real RDBMS.
Exactly.
The big guys are not really selling support but more of an insurance policy against failure. (whether they live up to that is another story).
one analogy, if your car stalls at home, sure you may be able to fix it, or your neighbor might help you out. But a large number of people buy AAA or some breakdown service, for when they can't fix it themselves or the neighbor isn't around. Its called peace of mind.
Oracle are the GM breakdown service, your neighbor who works at the local service station is your open source guy.
Yeah, but in practice, even with up-to-date versions, there are lots of bugs with the BDB tables (table format required for transaction support). When I switched my site from MyISAM tables to BDB tables in order to benefit from transaction, left joins stopped working, "show table status" core-dumped, etc...
MySQL : nice toy to play with, still not a real database.
Also Postgres doesn't support blob, you have to use a funky "large object interface" that plainly sucks... and unsigned integers are not available too... and it seems Postgres files are not very crash-proof.
...and it is very buggy (I tried : left joins stopped working as soon as my tables were converted to BDB tables)
I've been an Oracle DBA for more than 10 years, contracting for the last 5, and I can say all
the projects I've worked on have one thing in common - there're EXPENSIVE.
Without an expection, the cost of the support staff and developers has always exceeded the
the cost of the Oracle software and hardware combined.
If $100K is too expensive, then you probably can get by with a cheaper alternative. If you can
afford more than $500 k/year, then Oracle orDB2 (mainframe version) are probably the products
you want, with the features you need.
"The best part? I became an ordained minister while not wearing pants." -- CleverNickName
I work on reservation systems and trust me the volume that we get from the web is nothing compared to the volume from other sources.
fyi: we (and most of the large hotel chains) use Informix, it rocks.
however for really really really big databases nothing can touch Teradata. It's not fancy but is really impressive on the performance we've gotten from it.
Why do you think that you're qualified to speak on this matter?
From reading your comment you're just adding to the noise.
Have you ever deployed/worked on/maintained an enterprise scale database?
If not, then please shut up.
You can not do everything in "plain efficient PL/SQL"; and even when you can do it, it's simply reinventing the wheel. Try writing a millisecond-accuracy equivalent of the "SYSDATE" function in PL/SQL.
Oracle itself makes clear that it has never intended Java in the database to replace PL/SQL. Java on the database is here, it works very well, and it will be here for a long time to come. It's a pretty neat idea whose time has come.
Zigbee Central: A Zigbee weblog
Mysql is fine for a small number of users in a shop where data integrity isn't critical. But if your database is an enterprise database with strong data integrity requirement, you need something like oracle. It has stood the test of time for quality, for an enterprise scale operation. When you have hundreds or thousands of people relying on your database being accessable, that enormous oracle tax, (I always thought it ran more (a LOT more) than 10 to 20 thousand dollars) doesn't seem so excessive.
My Weblog
You've been in the consulting world too long, time to branch out.
Never go to sea with two chronometers; take one or three.
You should have mentioned Postgresql. This is a full powered rdbms that is open source! For all but the most heavily beat on application server environments I would put it up against Oracle or Informix any day.
MySQL, despite its advantages in certain delimited application spaces, is not a real RDBMS yet. When it gets full transaction support it will be closer. But it is playing catch up on some key RDBMS features.
This is FUD. Postgresql has beat the pants off of commercial dbms systems. Will it? Try it and see. Don't just assume it won't work. Mission critical apps? Do you know how to set up a test harness?
Agreed on MySql. Stored procedures are starting to be implemented directly in Java (about friggin' time!).
Are you implying that Postgresql does not have every single one of the things mentioned? BTW, clustering, if you mean the old-fashioned idea of storing data from multiple tables on the same data page, is a REALLY BAD IDEA that used to crash Oracle really often in the early days when they tried it. It was largely shown to be a bad idea back in the 70s. But it occassionally still rears its ugly little head.
This isn't exactly difficult to add. Is it such a killer feature for your work?
Using RDBMS-specific code is vastly important to getting real performance out of the database. Of course you can write everything using Perl::DBI and access the database with pure SQL, but is that really as quick as putting stored procedures inside the database? Alternatively, you could use an abstraction layer and hide the "native" stuff, but you still have to re-implement for each back-end.
Also, part of the value of the big players is the add-ons. Doesn't Oracle claim to give you enormous performance benefits for the web if you use their products end-to-end? Are you really going to write your own text indexing system if Oracle's ConText/InterMedia is available?
And even if you do keep your application portable, you still need to invest in administration skills. Whilst almost anyone can fumble around the SQLServer Enterprise Manager GUI, you still need specific knowledge to really get the most out of it. And developers who think they can stuff data into Oracle without considering the specifics of the database are setting themselves up for a fall (or perhaps a crash :-).
Who would believe in penguins,unless he had seen them? Conor O Brien - Across Three Oceans
If you want quality support from Oracle (or MS, or any software provider) you have to pay for it; the free support isn't. You can buy support for Linux, MySQL, PHP, etc. as well--so the support argument doesn't hold water.
an oracle tar is kinda like a bug report
sev 2 is the severity level
The actual link is here and it's not just for MS SQL Server. They specifically mention BEA and IBM, probably 2 of the biggest players.
They only will help tune the site if it "out of the box" does not run 3x faster.
I think its a hella fine claim, even if they do use it just to strip more licences and force their recommendations down your throat. With the big boys always fighting over the max throughput who wouldn't want to run your site 3x faster?
When searching for an opensource alternative for Oracle, I came across a frequently asked questions page on MySQL. One question asked "Is it stable?" and the answer was "MySQL has been tested under the numerous conditions.. blablabla" but the effective answer was they could not be sure it was stable in all conditions. The moment stability and therefor data integrity became an issue, my searching for an open source database was over. A DBMS should have 100% reliability, anything less and the system may as well be free. If you are unsure whether any of your data is incorrect, then you can consider the whole database corrupt. You pay for a system such as Oracle to ensure your database is stable.
http://github.com/gbook/nidb
I don't think /.'s "hiccups" are the result of MySQL problems, more like...
"'DELETE FROM tablename;' Oops, SHIT, I forgot the WHERE clause! Where's that damn backup?!?!"
Sticking feathers up your butt does not make you a chicken - Tyler Durden
We are currently running an iPlanet(JSP)/MSSQL-based solution, and are investigating other possible db-solutions. We are using a home-grown system for generating Java classes and sql-code for stored procedures and tables out of UML-models using Python and ObjectDomain.
What I've found to be the biggest problem with "free" dbms such as postgreSql and MySQL is the lack of stored procedure support and/or arbitrary naming of primary key fields. Switching to Oracle would also require quite a bit of work with re-writing the code-generation scripts, but not as much.
We will probably stick with MSSQL after all (going to MSSQL 2K), but not because of performance reasons, or high-end features like clustering or replication, but rather because changing to another dbms requires more work than it's worth. I believe this situation could be quite common.
If we were starting from scratch today on the other hand, we would probably have gone for a "free" solution, and designed the tools accordingly.
You say there's a guarantee that "If the Oracle database is setup and properly configured, YOU WILL NOT LOSE ANY DATA."
Really? Does anyone have a link or any other info about this? What is the exact guarantee?
I've always thought that liability would be a good product to sell in the free software marketplace. I never thought that it was a very insightful or original idea, but I never knew of anyone actually selling such a thing.
I'm going to have to eat part of another post I made if this is true... Oh well.
--Lawrence Lessig for Congress!
Is it worth it? Yes, if you take a more sociological as opposed to technological view of it.
If you use what others have used (and continue to use), then you're protecting your back against a failure. If you go out on a limb, use free stuff and it fails, YOU are reponsible (at least initially). If you do what others have done, then if it fails, at least it wasn't "your fault".
I did my M.Sc dissertation in a hospital, doing a CBR helpdesk for some life support equipment maintenance. In the end, practically nothing could be done because if something was done due to the AI's recommendations, then the programmer (and the supervisor who allowed the programmer in) would have to be responsible. If we left things as they were (and let the few overworked technicians do what my CBR helpdesk could have done), then its their responsibility (and everyone's happy).
Karma makes sense. It makes a lot more sense if you add reincarnation.
I talked with an admin for a medium sized ISP and he said that the best they could get mySQL to do with their customer dataqbase was 2 days and he stated that Oracale hadn't crashed their systems for over a year so I would imagine that is worth it instead of rebooting every day
Huh? Are you talking about the fill factor? If so, that has little to do with locking, that is used to decide how many rows to leave empty on each page, to leave room for insertion of new rows in that page at a later time. This is mostly important with clustered indices...
Sybase has supported row level locking for a while, in fact locks start out at the row level and then get upgraded as needed to a page lock, then an extent lock, then a table lock, if memory serves correctly...
philg waxes poetic on databases.
--- this space intentionally left blank.
...and just call their help desk and mention 'Perl' and see how far you get.
Peace, or Not?
For those readers without much database experience, Philip Greenspun has written a readable introduction to databases and RDBMSes.
IE Oracle Applications, SAP or PeopleSoft... These are the major reasons that Fortune 500 use the Oracle of the world... Not to as a backend to their website...
And remember that most Fortune 500 companys are not in the database or software business, and it would way outstrap their IT dept. to attempt to support custom apps on an non-commericially supported backend....
But, if your making /.2 in your basement, you are crazy to use anything other than an open-source rdbms for cost alone...
--
--
Poliglut.org: 75 Million gun owners can't be wrong
Keep in mind that prefacing something you just made up with "as the saying goes" does not really do any good.
No one ever got fired for buying I.B.M.
Of course everyone can't keep him around for their own database, but you could pay employees to learn how to use the product for less than Oracle support
Maybe he misspelled "Whitey."
Do you realize you misspelled misspelled?
Even if I already have an Oracle8 DB? I can smell easy money here. :-)
Within the past 6 months, I took a new job, leaving behind Oracle, for MySQL. I also packed away my $300 worth of Oracle books, and replaced them with one book on MySQL.
There are many things I now miss about Oracle. MySQL does allow user functions, but they must be written in C (instead of PL/SQL), and they need to be recompiled for each environment. MySQL doesn't have the useful ON DELETE CASCADE, or any of the features Oracle gives you for foriegn references. MySQL's procedure for creating unique ids is to set AUTO_INCREMENT for a field - much simpler and easier than Oracle's procedure: create a SEQUENCE, then create a TRIGGER ON INSERT to pull the next number from the sequence and insert it as the id. The Oracle version is a lot more complicated, but offers a much better functionality, functionality that I wish I could use in MySQL - more than one table in Oracle can use the same sequence! (This means you can join multiple tables, and they can still have unique ids). Oracle also allows you to create a VIEW - kind of a virtual table based on a select statement - which other developers can then use as if it were a table. And probably the most important to it's users, Oracle has a multitude of very decent front-end tools for administration as well as simple data access.
If you (and your staff) have a larger checkbook than development & database experience, you may be better off investing in Oracle, Oracle-related books, and maybe even some Oracle classes. It's arguably the best relational database that money can buy. There are also a ton of books, and classes available to bring your staff up to speed with using Oracle.
But being in a shop full of developers, we can get around what isn't there on MySQL, and do it for far cheaper (including the cost of our time) than purchasing an Oracle installation
As far as support, if you think MySQL is unsupported, you better read their licensing policy. I've found they actually have very good support - that is, after all, where they make their money!
The next generation search engine -- TRY IT!
"The large print giveth, and the small print taketh away" -- "Step Right Up", Tom Waits
I use Progress at work. We have a few big business critical systems based on it (including MFGpro). It's at least as good as Oracle but a lot cheaper, and the 4GL is a general purpose programming language. Progress have also made an effort to make web integration easier with their WebSpeed product.
However, enough of that.
I also use PostgreSQL 7.0 for some development work. If you need a fully-functional RDBMS and free licensing is an issue, you will be very happy with it. And since there are Python libraries for it and it works well with Zope, there's nothing I can think of that you couldn't do with it.
Ask yourself: do I want to have to buy licenses before I get started and keep them up to date and legal throughout the lifecycle of the project? If you're spending the company's money rather than your own, it's not just about cost. It's about being able to move fast and being free to explore and experiment.
-- What do you need?
-- Gnus. Lots of Gnus.
Check out the TPC benchmarks (a standardized set of very strict performance tests) for your answer. Pick either by raw performance or by price/performance. You don't have to be a genius to understand the numbers: http://www.tpc.org/
Did you even READ the article? They're using it for some cheesy portion of a website -- for dissemination of information. I don't know what they use, but I can tell you this: they won't stake the lives of astronauts and billion dollar satellites on MySQL.
I have worked for two companies during their migration to oracle. One started with PostgresSQL and the other MySQL.
Neither database system was keeping up with the load. Certain large, unfriendly queries would cause the database server to crawl or even crash.
None of the open source databases have complete funtionality. Some of them are starting to get more in recent (I mean very recent) releases, but they are still not going to have the stability you need when you company depends on the database for all of its revenue.
SQL Server is all written by Microsoft. Pre-version 6.5 there was some Sybase code it was based on but what relevance is that? SQL 7 and SQL 2000 are their own breed. See http://www.tpc.org for 3rd party benchmarks.
While all you blowhards are trumpeting Oracle, you might consider the cost and performance leaders. See http://www.tpc.org for a constantly updated list.
http://www.tpc.org/tpcc/results/tpcc_perf_results. asp offers some answers. Of course few folks who read slashdot actually care about which database might actually be best. Instead they blindly claim JimmyBob's Free DB with Source is best followed closely by the Mr. Larry Ellison's Oracle. Fine. Sign a deal with the devil.
I think you hit the proverbial nail on its proverbial head with your last sentence:
In many --even most-- cases, the high-availability and extraordinary scalability of Oracle aren't only not necessary but downright overkill. This is true because most of us don't build air traffic control systems, massive financial services transactional systems, and the like. You don't have to know much about databases to realize that at 3TB with 10,000 transactions per minute, Oracle will outperform PostgreSQL. At this level of activity, the Oracle license is most likely considered nominal to whoever is footing the bill for the project. However, at 4GB with 10,000 transactions per hour, you might find the runtime difference between PostgreSQL and Oracle not enough to justify the expense.Incidentally, don't be fooled by the mythical status of Oracle (the database) and Oracle (the company):
But to get mired in this debate is not only to miss the point, but to engage in the sort of futility that causes projects to fail. Smart projects minimize technology risk. In general, the best way to do this is to build systems that are platform-agnostic.
Too many project managers/sponsors (especially in the enterprise) are engaged in the delusion that their project actually warrants a massive Oracle cluster, four-way Sun Enterprise E10000's, three Oracle DBAs, a posse of Java/EJB consultants, 30 seats of PVCS, 10 seats of JBuilder enterprise, and a partridge in a pear tree. Ask this captain of information technology what he's assembled this crack team to do and he'll likely respond with 'A custom CRM solution', or some such. This, of course, is stupid, and what's worse it's bad business. Naturally no regular reader of Slashdot would engage in such poor behavior.
The opposite of this project is the smart IT guy who realizes that the business (or marketing department for commercial developers) never knows what they want, how big the user base will be, or even what kind of cream cheese to put on their bagel. To compensate, he starts building a solution using the lowest-cost solutions that will still allow him to scale. In the case of the database, there is no reason NOT to use PostgreSQL with ANSI-SQL through a database-independent interface (such as jdbc) during development. In some cases, it may even be appropriate to continue to use PostgreSQL until the production environment warrants conversion to a more scalable, available solution.
This philosophy applies equally to hardware, tools, third-party libraries, configuration management, or even IT talent. If you're looking at a new project and someone is telling you that you need a bunch of high-priced licenses and assets at the outset, examine the rationale behind these claims carefully, chances are you don't (yet).
Art Gillespie
Art Gillespie
J2EE Consultant and all-around nice guy
actually, just for the record, (cuz i've seen this twice so far, and am only halfway down the posts), mysql as of now *does* support transactions. check their website on the most current version for more details.
wow, M$ Sql server on sparc? how'd you manage that?
To say that a website is not mission critical is dismissive of online companies such as Amazon, or the fact that online ordering and purchasing should have transactional capabilites, and triggers in their DB. For sites who seriously want to be involved in the ECommerce, you might need a bigger DB then mySQL. mySQL is great for storing/serving content, but adoption of mySQL for large ECommerce sites should be considered carefully.
Actually, the question is a good one, even if it does inflame the opions of the ignorant and opinionated. Many, many new people are moving to Linux, and many of them are old-time (or even young) DOS/Win* hands yearning to breath free or whatever. Most are used to small, readily useable (I say this with tongue in cheek concerning Access) dbms's that permit a single user on a single machine to work with relatively small databases. When these people first encounter Linux, where are those small, easily used database systems? If you are an ex-DOS/dBASE/Paradox hand something like the commandline of postgreSQL may be a breath of fresh air after being in prison for 10 years, but for most, the dandruff from head-scratching piles up in drifts on the desktop. Even some flaming opinions are better, more current, and easier to execute than the search for information buried in the docs entombed under /usr/share/docs/HOWTO/html/en. --- Theories, they'ld map out roads through hell with their crackpot theories. - Dragon to Grendel JWD
------ The only greater hazard to your liberty than n politicians is n+1 politicians.
You can't Sybase's off of a free version of their commercial database.
Free for commerical use version of Sybase 11.0.3 it is available at http://linux.sybase.com. Note that the other version are pay but their prices are reasonable (not like Oracle).
Free JDBC for 11.0.3 is here: http://www.sybase.com/products/eaimiddleware/jconn ectforjdbc
blessings,
"Only in their dreams can men truly be free 'twas always thus, and always thus will be."
--Tom Schulman
Oracle on NT is a joke in and by itself
So where does Oracle shine ?
Sacalability. 4- 8- 16-way high horsepower servers with terabytes of data.
Distributed databases
High Avalability setups
featues and tuneability
But, having said that, there are enough applications out there in the wild that are so horrendously bad that they suck even with the most advanced database under them. Cisco Secure comes to mind....
f.
I have two main problems with MS-SQL:
DB2, Sybase, and Oracle all provide cross-platform support, the ability to use a CLI admin interface which facilitates scripting, and cross-platform support. (I list it twice, because for me it is very important).
With MSFT, there is a real risk of vendor-lock -- they aren't just selling the RDBMS, they're selling the OS and the apps. I've found it to be much more difficult to get MS-SQL to play nice in a multi-platform environment, particularly when it's not MS-SQL that's the "main" system.
MSFT's claims of scalability are highly exaggerated. While it may run decently once the data are loaded, indexed, etc., the real limitations of MS-SQL become apparent while doing the loading, indexing, etc., as well as the day-to-day data management tasks. Their tools are inadequate at best. I don't think MSFT yet understands data processing.
Gordon.
He that breaks a thing to find out what it is has left the path of wisdom.
-- J.R.R. Tolkien
Actually you can (for now) access MS-SQL from *nix using PHP, Perl, Java, etc. Use the Sybase Open Client libraries, which you can d/l free from sybase.com.
Once MSFT drops TDS (Tabular Data Stream), then we would have to look for other middleware. Unless, of course, we've migrated the heck away from MS-SQL. Sybase would be the easiest target, but I'm not sure of their ongoing commitment to Adaptive Server Enterprise. Thus, my migrations are to DB2, often on Linux.
I'm assuming that the statement "none of these databases can comete with Oracle head to head" applies to the open-source databases. I'd be most happy to put DB2 up against Oracle (and Oracle up against a wall. ;-)
Gordon.
He that breaks a thing to find out what it is has left the path of wisdom.
-- J.R.R. Tolkien
I'm not guessing, I'm uncertain. Yes, I see them rolling out ASE 12. Very good. But I don't see them marketing it in any visible way. They have been marketing their portal and, to a lesser extent, EAServer.
I like ASE. A lot. I have clients who happily and productively use it. At the same time, I don't see Sybase marketing it. It's a tough sell when clients haven't even heard of it, but have heard (and heard) of Oracle.
Here's an example. The local Barnes&Noble has an entire section of Oracle books. Two, maybe three titles for Sybase ASE. Now, you and I might look at that and say "Dang! That Oracle must be really tough to use.". But most people (in my experience) instead see that as a sign that Oracle is widely used and can do a lot, whereas Sybase doesn't seem to have much demand. Oracle has captured mindshare. Sybase hasn't. I hope they start.
On the other hand, I know DB2 will be around, because I see IBM aggressively developing and marketing it. That's a much easier sell.
Gordon.
He that breaks a thing to find out what it is has left the path of wisdom.
-- J.R.R. Tolkien
Just like that old phrase says, noone ever got fired for buying Microsoft.
Just a historical point, the actual phrase is, "Noone ever got fired for buying IBM".
I know of at least one guy who got fired for buying Microsoft, and a few more (including my predecessor) who should have.
Goofy, Geeky Gifts and More!
If Oracle is so great.. why did NASA switch to MySQL ?
UPS Sucks
I'm working on a Web application currently running on MySQL and I am interested in having work with any database (Perl/DBI). One feature that is especially nice when searching the database is using LIMIT in the select statement so that I don't retrieve 50,000 rows when someone searchs for "the". Does Oracle have anything like that? From what I've seen the answer is no.
I thought that the 1,993 revision number jump was pretty impressive... (7 -> 2000) ;-)
--
News for geeks in Austin: www.geekaustin.org
News for Geeks in Austin, TX
Umm, have you ever heard of a company suing a software company over a computer crash? I mean specificly because of a crash. and not due to breach of support contract.
Something tells me that it's neigh impossible to sue over software. 8-P
So my opinion is that choosing a brand name cause you *THINK* they have money to go after if it falls apart is not a valid reason except in PHB's eyes.
What? me have a sig? don't be ridiculous.
Think about the Avis, American Airlines, or Best Western websites. Don't you think those sites are mission critical and hold vast amounts of data?
I would guess that the databases "attached" to these web sites don't hold much information at all. They hold some front-end stuff, maybe names and information about locations. But they undoubtedly connect to back-end legacy systems that hold reservation information, which would be the higher-volume data. These are often kept in old mainframe databases that aren't relational. Think about it, do you think that Avis wired up their reservation system (and all of the travel agents who use it) to the db for their website when they made their website, or do you think they wired the website to use their pre-existing back end?
I've had an opportunity to help develop quite a few RDBMS based applications for various customers ranging from the backbone for a distributed chat system to NSI's BARS (Billing And Receivables System) and every single one of them was implemented using Oracle. Why? Our customers run systems that see millions of transactions in a day -- something that most other RDMBS's out there simply can't handle. Oracle is reliable and is a hell of a lot more scalable than SQLServer, but that's beside the point: our customers use Oracle because they trust it. Would you implement a billing system that doesn't implement some kind of transactional integrity? Most sane people would answer no, because they understand that 'low cost' solutions are not necessarily inexpensive.
For serious transaction-based endevours, MySQL just doesn't foot the bill. This is for one reason, mySQL does not support transactions. _Desmond
_Desmond
You have one hour to answer the questions. Write legibly in black ink. If you require more paper raise your hand and a member of staff will supply you with some.
^^^Mod Up Funny^^^
-- the computer doesn't want any beer, no matter how much you think it does. NEVER, EVER feed your computer beer.
Do you mean it's there where http://ninenine.com (No bullshit, no popups, 100% free porn added (twice) daily! NineNine.com) gets its pictures?
Well, I suppose it makes sense. Those "extreme vaginal close-ups" MUST come from some sort of "medical" database, right?
A man of this class cannot be wrong
--
Je t'aime Stéphanie
I am not qualified to comment on Oracle or other RDBMS'es but I would vouch for MySQL paid support. They are very quick to respond and work until the problem is solved. It's worth the money.
I gaurantee that Oracle support won't get you email from the main developers. When you have MySQL support you may just get Monty to respond.
If need be they'll send a patch for you to compile, so you can see what changed. Oracle won't be doing that anytime soon.
Oracle seems to be pretty confident that their database servers will be 3 times faster. I see their ads for a $1,000,000 guarantee that Oracle 8 will 3x be faster than your existing server.
Are you on crack???!!!!
Transactions are there so you don't *have* to implement that sort of thing in code.
That's the difference between an RDBMS and some garbage like MySQL.
Trust me. They are needed.
> You mean you _ever_ would have?
How do you think he knows about the MC smartcards?
sigh
MySql is a nice little database. I love it for small websites that I do on the side. I also use it for various things at home on Linux. At work I'm an Oracle DBA. I would not consider MySql for anything "large", in fact I really can't imagine using it for a high volume website with thousands of transactions an hour. Can you say data corruption? How about file contention? I won't even go into the features it lacks compared to an enterprise class RDBS. Its not made for that and I don't think anyone expects it to be. Now you mentioned Access... Just use Excel, it takes less mouse clicks. As for MS Sql Server. Well, its a very nice platform to develop on and a pretty good database to boot. Don't expect to put any data warehouses on it and don't think you're going to have thousands of concurrent sessions on it either. That is not a short coming of the RDBS (Sybase has been doing it for years), its the operating system. NT just isn't scalable. If it was, Microsoft might finally be able to transfer Hotmail or their own website onto it :) Hell, they might even chance putting their own critical data on it.
If M$ were to port SQL Server to Unix, Oracle would really have some competition.
The fact is, all RDBS's have their uses. If you need performance, reliability and scalability and a lot of money then buy Oracle, or one of many others that aren't as feature rich, but very good systems - Sybase, Informix, DB2 or MS SQL Server.
The only advantage I really see with those is potentially brand name, and support. What I mean by Brand Name is chances are your boss has heard about Oracle, and heard only good things about Oracle, but never heard of MySQL or only heard a tad about it being Open Source. And then there is support. If you are going to run some kind of eStore with potentially millions of customers and transactions and products, and then one day your PHP and MySQL setup craps out and you have no idea why or how to fix it in a short amount of time, who exactly are you going to turn too? You have an entire business, employees, and customers that are riding on your database setup, what are you going to do? You can get MySQL support, and sorta support through Zend for PHP, but I know their support staff is nothing compared to Oracle, and I know with Oracle you can pay them to actually fix the problem and not just give advance.... it all depends, how mission critical is your database?
..There's a-dooin's a-transpirin'
Who are you going to sue if MySql goes down for a day?
Um...nobody. The exact same thing will happen if your Oracle database goes down for a day. Please don't tell me you honestly think you could win a lawsuit against Oracle for lost business if their database went down for a day? There's just no way. Read your EULA. I'm sure it excludes such things.
The smart thing to do is to take the money you'd use to hire the lawyers to sue someone, and instead pump it into planning and redundancy and disaster recovery procedures. If you don't take precautions like this, then you might as well give up now, because it won't be long before all your database are belong to us.
If you'll be blamed if it breaks, then you need Oracle.
If someone else (or noone) will be blamed, then anything else will do.
Maybe the state's highest function is to grind out insoluble problems. (Zelazny, Hall of Mirrors)
http://otn.oracle.com
you will need to register, but once you do you can download developer versions of their software, including IFS 1.1, which maps a filesystem to tables and lets you share that data via ftp, http, imap, and ntfs.
i'm just getting started with oracle. it's complex and resource intensive, but from what i've seen so far, it's a top notch product and has excellent administration tools.
make sure you have plenty of RAM and disk space before you even think of installing it. the oracle service for a single database takes up around 85mb of RAM and at least 500mb
~pixel
btw, when will slashdot start supporting XHTML markup when posting? <p /> isn't in the list of allowed HTML tags :(
And when I first heard it, it was Cisco.
I'm sure there are several variants.
Should/Would anyone have any confidence in this algo until it had been widely analyzed for many years? No.
It is the same with databases. You might test a huge PostGreSQL db for years and get some amount of confidence in its reliability and scalability. But until it has been tortured as much as Oracle, it won't be trusted as much.
Cautious DBAs will be slow to accept free databases, just as cautious cryptographers will be slow to accept AES (Rijndael). There are already more time-tested alternatives like IDEA(crypto) and Oracle(db). It is not that they distrust the newcomer, per se. It is just that the stakes are so high that the utmost caution is demanded.
In databases and cryptography, there is only black and shades of gray -- either proven unsuitable for a task; or so far, so good. It is almost impossible to prove stability or security.
You can never equivocate too much.
Well, speaking as a DBA with 12+ years experience, I would have to disagree. I was a Solaris fan for many, many years, but lately AIX has won me over with it's file-system management tools. It is dead easy to dynamically re-allocate your physical storage in AIX. If you've never played with it, do yourself a favor and try it. smit is your friend :-)
Why is it that the proponents of "one nation under God" are so eager to get rid of "liberty and justice for all"?
What is the Open Source world's obsession with MySQL, and wanting to use it to back-end EVERYTHING? I use OS in a corporate enviroment and there are lots of packages I can't use because of this. This is a little bit of a rant, but if your going to develope RDBMS backed applications please take the time to learn a little about database systems, of which MySQL barely qualifies. Anyway, applications should use ODBC, or better, gnome-db, so that a user can use whatever RDBMS they have (PostgreSQL, Informix, Orable, MS-SQL). Also M$-SQL doesn't cost $10K~30K? When did that happen? Comparing Access or MySQL to Oracle, etc..., simply reveals phenominal ignorance.
Using "Common Sense" is being either to arrogant or to ignorant to ask people who know more about something than you.
You can fix it yourself if you or someone on staff knows the source code well enough to be trusted. Access to source is a widely touted virtue of open source, but I wonder about the real world impact. And, yes, you can pay an outside vendor to provide support. But, it seems to me that one key difference between closed and open source vendors is that the reputation of the closed source vendor is at stake if they can't make their code work. If a third-party support provider can't fathom, say, MySQL, what's to stop them from walking away?
-- Slashdot: When Public Access TV Says "No"
...yes.
You are very right on this. You have to think much larger than a website for this question. When you set up a truly enterprise scaled project using Oracle, there is not much room for the word unproven. And even if Mysql could manage databases which are several Terrabytes in size, what about the interfaces and development solutions? I am thinking Java for Oracle, Oracle Applications, etc. Most people who don't work with large scale servers or mainframes seem to think in what I call "The PC line of thought".
They imagine large amounts of transactions (for them) to be something that Oracle DBA's would laugh at.
-- Who slipped something in my coffee this morning? (And can I have some more?)
Backup & Recovery: THIS is the reason people pay money for Oracle. You'll never understand how amazing oracle is until you've LOST data and have to restore it. The ability to restore a datafile without taking the database down is a wonderful thing...
that's it.
If your open source screws up everything badly you have no one to blame.
If your MS/ORACLE/WHATEVER does the same, at least you can yell at them on the phone.
Most of my work involves programming against Oracle Databases, primarily in Java. While I used the sqlj translator for the first major project, I chose not to from then on. Mostly due to JDBC being rather well implemented, and the fact that by using JDBC directly, I can eliminate the runtime components and cut down on the memory footprint of my apps.
As far as PL/SQL goes, my preference is for Java functions in Oracle 8i. From my background, it's just more efficient for me to work this angle. Only major beef I have with Oracle with these is the thread model they use inside Oracle is not pre-emptive. Have to be careful.
In contrast to all the the other databases you name, Access is not a database server, it's a file format. Any searching and processing of the data is done on the client side, not the server side.
So, the computer running your Web server would have to run the Access (aka Jet) software too. With the others you have the option of offloading the database to a different box and just transmitting the results over your network back to the web server.
Access also has an absolute upper limit of 255 simultaneous users. Just how many hits are you hoping for on this web site of yours?
Microsoft themselves are moving away from Jet for even lightweight standalone database work. Instead, they now have an "SQL Server Lite" - the Microsoft Database Engine (MSDE).
I deal a lot with backups, ARCServe, BackupEXEC and a bit of NetBackup.
The point I am trying to get to is that for most Comercial Backup Software you can find MSSQL, Oracle, DB2, Informix, Sybase agent to do live backups.
Yes I know you can do a dump and backup up the file system. How often do you see that being done (with the possible exception of MSSQL 'cause restoring through an agent is at best painful)
Yes I suppose you could get out the PERL DB modules and home brew your own "agent" to do live backups, but who's gonna pay for that?
I guess the point I eventually going to arrive at is that it's not necessisarly the support the vendor (or developers) provide it's the support the rest of the industry provides. What do other large enterprise software firms take seriously?
Having said that I truly like mySQL and have "sold" it to several people, but you can't compare mySQL to DB2. Or at lease not now, in a couple of months (year?) who knows?
When someone yells "Stop" or goes limp, or taps out, the fight is over.
Ya for that!!!!
When someone yells "Stop" or goes limp, or taps out, the fight is over.
Oh, no. It's dumb.
This is an optimiser that decides that table scanning a table with 3 million rows, is preferable to scanning one with 100k rows.
That to me, is dumb.
I guess it depends on what you mean by "break": do you mean they don't work at all, return different results, or the optimizer just picks the wrong index?
I mean it picks the wrong index, runs for hours and fills up the tempdb.
That's breaking.
We had all our thousands of procs tweaked to run perfectly under Sybase11, and they changed the optimiser so much that it now makes such stupid decisions that we have to forceplan about half our procs just to get them to run again.
That's pathetic.
11.9.2 was a big step forward, especially since it lets you force the use of an index, but there's still room for improvement
I'm not sure what they did in 11.9.2 to help this.
You could index hint in 11.0, but it still doesn't get the optimiser to make decent choices.
We are really relying on forceplan to get query plans that make sense.
We shouldn't have to do that. The point of the optimser is that it should, in the general case, get the query plan right without index hints, and without forceplans. For our app, under sybase12, that is simply not the case.
--
Read more of this story at Slashdot.Read more of this story at Slashdot.Read more of this story at Slashdot.
Stop guessing, and ask them.
ASE is still a core product of Sybase, and will be for a long time to come.
They are putting a lot into ASA, but it's not at the expense of ASE. They've cut back on things like PB, and diverted resources from there into the mobile market.
Sybase is very strongly committed to ASE, and they're releasing newer, better, faster versions, at an accelerated pace. ASE 12.5 is in beta.
I would trust my data in Sybase, without a worry.
--
Read more of this story at Slashdot.Read more of this story at Slashdot.Read more of this story at Slashdot.
Sybase is still working in the same market. ASE is now, feature-for-feature much closer to Oracle than it ever has been.
Sybase's core strengths are still there (it's so much easier to use/manage), and the high end features that people have been demanding are improving.
Sybase is a big company.
They are pushing into a number of markets, but they're all related. Their big push is into the mobile database market, and they're doing incredibly well (they have a 60% share).
That doesn't mean that they're abandoning ASE. Sybase has always had interests in a number of products, notably powerbuilder. They're shifting their peripheral efforts sucessfully into them mobile market, but ASE is and will continue to be a core product.
I have no concerns about the future of ASE.
--
Read more of this story at Slashdot.Read more of this story at Slashdot.Read more of this story at Slashdot.
Oracle beats Sybase with:
Dammit guys, can we please, please, please have them now!
My other issues with Sybase are:
They re-wrote it for 11.9/12, and it broke half our procs!
It needs to support more implied conditions. Transitive joins are good, but there's a lot more they could do. Particularly relating to indexing on low precision datatypes.
I'm quite happy with SQSh, but some of our developers like GUIs, and Sybase Central is pretty poor, and DBArtisan is a complete joke.
But in terms of bang/buck, Sybase gets my vote.
--
Read more of this story at Slashdot.Read more of this story at Slashdot.Read more of this story at Slashdot.
If you read the scores more closely, all the MS-SQL are clustered solutions, while the best single-machine performance is Oracle 8i.
Since throwing more machines at a problem will eventually get you more performance in many cases, it would suggest that MSSQL can't compete per 1 piece of hardware, and that the Oracle folks never bothered to put their software to the test on a "real" cluster.
--eric
More data, damnit!
Oracle and other high-end databases get a great deal more lab testing than open source alternatives do. They are also tested in environments that open source testers wouldn't be in. Want to test a billion transactions today? Let's use Amazon's inventory...
Also, database companies that charge outrageous fees for their software tend to hire the best database programmers because they can afford them because they charge outrageous fees for their software...
And never forget accountability. If my database crashes, I WANT someone to point at and say, "FIX YOUR DAMN PRODUCT!" With open source, I don't have that.
Having said all that, open source databases are great for many, even most projects. But Oracle and the like are best used when you need all the bells and whistles and most importantly, someone to blame other than yourself for implementing it. "But boss, I swear MySQL works great, most of the time..." No thanks. I'll take Oracle and a scapegoat.
Why do users with IDs under 100,000 or over 700,000 usually have the most worthwhile comments?
No, you're paying for the features.
:)
Service and support costs extra
"There is no night so forlorn, no mood so bleak, that it cannot be infused with pleasure by tender meat..." - R.W. Apple
Offtopic?!? Ugh. I daresay it's funny *and* insightful.
There's a reason why UPS, FedEx and such large companies have databases (as in: RDBMS software, mainframe computers, and large array of drives) that costs $$$$.
Just because it's cheaper and open-source, it doesn't mean it's the better solution -- it's foolishness! (Of course, using Teradata for storing your company roster is foolishness in the other direction.)
Sigh.
as of yesterday, one of the instances of the Walmart database was 29TB.. just over what that one guy said was the largest in the world.. but you are correct in saying that the data, for the main part, is spread out over a bunch of databases. Working at the data center is a lot of fun... trust me... (never go home when there is a problem = loads of fun)
Jay
"What's this script do? unzip ; touch ; finger ; mount ; gasp ; yes ; umount ; sleep Hint for the answer: not everyth
I have extensively used both of the Microsoft Database products Access and SQL Server. I have found that there is a huge difference between the two. (One interesting side note, M$ claimed that Access was a full competitor to Oracle until it released SQL Server at which point Access suddenly became suited only for the desktop - which I might add is where it belonged all along)
Access uses a Jet engine which is not a relational engine at all, while it supports many of the relational semantics - the way it handles the calls under the hood is not relational at all. To get a result set from a query, it locks the entire table and loops through it one record at a time and examines each record, if the record meets the criterea it is then added to the resultset. Now imagine adding multi-table joins to this situation and you have multiple tables locked exclusively at the table level for a single query. This is fine for small (very) low user situations. However when you start talking about more than a handful of users you start getting serious table contention.
SQL Server on the other hand implements a very nice relational engine that is very fast and scales very well to large numbers of users. SQL Server offers things like indexing, B-Tree searches that operate on those indexes, a very good query optimizer, stored proc caching, connection management, row level locking and an entire host of other things including security, replication, clustering fail over support and many others that Access doesn't.
This doesn't make Access necessarily a bad thing, unless you are trying to use it for something it was never really designed to do.
I operate a small consulting company that builds web based database systems. In the past SQL Sever has been a nice alternative to say Oracle or Sybase because of it's relatively low price yet high performance and the abundance of features offered. However the trend of late has been for the price to continue to rise dramatically in the cost of SQL Server. I now find myself looking for alternatives such as MySQL that can eventually replace SQL Server. I've only played with it a little bit, but from what I understand it still lacks some of the basic features that will really set it apart as an Enterprise Class RDBMS. Simple things like transaction support and views are very important in many business applications.
I know that myself and many other small consulting companies are eagerly looking for a low cost (free) alternative to one of the more expensive systems, but as of yet I haven't seen anything that will work reliably enough and fast enough to do the job.
Doug Tolton
"The destruction of a value which is, will not bring value to that which isn't." -John Galt
When David Axmark, one of the creators of MySQl, gave a talk in Denmark in December he commented on MySQL versus Oracle.
He said you should use MySQL when you wanted a simple solution to your problem. You should never try to use MySQL as an enterprise solution. If you wanted 24x7 availability with no margin for error you should absolutely use an Oracle database. While MySQL would threaten smaller commercial databases in the near future there would be no contest when it comes to Oracle for at least the next 5 years.
The real difference between the Oracle's and the MySQL's of the world is investor confidence.
This is, of course, from the perspective of someone who has worked with quite a few start-ups looking for VC dollars. Whether the database was 20 MB or 1.5 TB, whether there were 50 transactions a day or 50 a second didn't matter. It was someone else's money they were spending trying to secure even more of yet another party's money. Why do affordable and adequate when you can do exorbitant and ostentatious? VC's see MySQL and say "huh?" - they see Oracle and say "These guys have their act together; let's give them a whole bunch of money so that they may sell monocles for myopic rabid monkeys."
Fortunately, for established companies, the dog and pony method of RDBMS selection isn't as universally applicable. Granted, they still have an obligation to impress investors, board members, etc., but they also usually have sufficient flexibility to make decisions based (somewhat) on needs and costs analyses.
Generally, tech wisdom holds that most companies have one good idea; everything else they make is mediocre. Oracle makes probably the best all-around DB out there; most of the rest just sucks.
Aside from that, you're completely correct. 'Keeping it portable' is a practical impossibility for a high perf DB; and the investments in time & people should really keep you stuck to your vendor for 5-10 years.
Different thread. Move along.
Seriously. Visualize 10TB DBs in GemStone. Do that, then we'll talk.
Oh, and see if you can get those OODBMS people to grasp querying rather than navigation.
Hey, this is easy. It depends!
If your running a bank, you want Oracle. MySQL is really unsuitable for anything that involves the exchange of money.
If your serving up informational web pages, MySQL is fine, since dropped connections and other network related errors will be a much more prevalent problem that database related errors.
For in-between problems it's a judgement call. I do think a lot of enterprises drop a whole lot of money on Oracle when it really isn't prudent.
RM
Adding the missing features that most companies need, and missing from OpenSource DB's would cost much more in just labor for the first year.
In my VERY bad experience with Oracle, I would say fsck em. Not only did it take over 2 days trying to get the damn thing to work correctly. It messed up my system pretty bad. ... for a mer $30,000 a year ...
When they have entire sites showing you how to install a single app, and its 30 pages for each install. Something is truley wrong.
Sure, they give you support
Stick with MySQL or better yet, PostgreSQL.
until (succeed) try { again(); }
until (succeed) try { again(); }
What do you get? * Scalability * Reliability * Maintainability And if the design is right, you likely get a little higher data integrity and complex rollback features.
- In a knowledge based industry your main asset will always be people -
I don't underestimate IBM... but I also think a Sun E10K ranks right up there with the E/390s. Oracle scales to the terrabyte range (albeit painfully)... I've worked with a lot of different kinds of systems, and I haven't found one yet where we can't make Oracle work (note the phrase - not that it's always our choice, but sometimes you use what the customer already has an enterprise license for :-).
Someday, I'll probably get handed a DB2 CD set and told to make a system work with it. As long as it talks SQL, we can do this.
I love vegetarians - some of my favorite foods are vegetarians.
I personally manage a Data Warehouse that I'm not sure could be done with MySQL or PostgreSQL. While both are nice products, neither one really allows me to maintain a database in excess of 1TB. While I know they can do it, I'm not sure its going to continue to scale the way we need it to. I agree with the others though that support is nice (I've used it several times), and scalability really does set the commercial products above.
In all reality I crashed Oracle 8i last thursday, granted it does take some work, but unless you're running the most current release 8.1.6.3.0 or 8.1.7.x.x there are hundreds of ways to crash it if you hunt around Oracle's support site. Most users will never hit across them, but some will.
Investors are more likely to drop money on companies using stable, standard software (like oracle) than those using Postgres, or MySql. Also, if you're selling a product with a database engine bundled, people are more likely going to buy from you if you can drop the "Oracle" name. Or maybe not.
That info seems a bit outdated- It specifies that Oracle can use roles to define user priveleges? I've never used mySQL, but I'm pretty sure it can do the same thing...
Eat. Now.
"I never metadata I didn't like."
Not sure if MS can replicate this kind of functionailty, but Sysbase can't (yet). If you need absolute performance and reliability, Oracle is the only choice (to the best of my knowledge).
there are no stupid questions, but there are a lot of inquisitive idiots
However, it still lacks:
Don't get me wrong. I like MySQL and I use it all the time. But if you're building a commercial site, you need commercial-grade tools.
Nuff said
Having worked with Oracle support I can say that they are very good. We ran into a problem that required patching and they had the patch to us in like 4 hours (we called at 10pm ona sunday night)
Oh, yeah, and don't forget the little thing called transactions. If you are writing a serious system without them, you will have a big problem. Last time I checked, MySQL did not support them, but PostgreSQL did.
...there's yer answer, cowboy.
I had a similar experience. I found a bug in the MySQL query optimizer which was causing our application to return incorrect results. I spent a couple of hours developing a simple test case and posted it to the mailing list. This was on a Saturday. By Sunday night Monty had replied to me in email, confirmed the bug, fixed it, and posted updated binaries on the web-site.
Had I been dealing with Microsoft or Oracle, there is absolutely no way that I would have had a patch available to me that quickly, let alone during a weekend.
This is not to say that all such problems get addressed as quickly in MySQL or any other Open Source package, but my experience has been that what they lack in front-line support they more than make up for on the back nine.
I've found that most front-line support problems tend to be documentation issues (i.e. poor or absent documentation, or a failure on the part of the user to find or understand the necessary bits). As a techie I can usually find my way around such problems, either by careful RTFMing, experimentation, and/or talking to other users.
Those approaches won't work with problems which are integral to the product itself. So if I had to choose between good front-end support and fast bug fixes, I'd take the latter.
-deane
Gooroos Software: plugging you in to Maya
-deane
Actually I was talking about someone who's an 'IT professional' but has no knowledge at all of what they are trying to support. I spent an hour on the phone to a guy who knew no Unix and no DBA skills - it should've taken 5 minutes.
But that's the extreme.
Our support contracts are based on the assumption of a suitably qualified and trained (possibly self-trained) DBA being available. Why? Because we're not supposed to be training over the phone, neither are we remote DBAs for that site. Both of these are services we provide but not via phone support.
Having said that we rarely enforce this. If a customer has a system down he really doesn't deserve to be told that he isn't qualified enough to speak to us. But once it's back up I might recommend a training course or spending time in the manuals.
A little honesty goes a long way. If I get a call which begins "Look I really don't know anything about this but the DBA's on holiday and the MD's on my back" - I'll be much more inclined to be forgiving than to Mr Nasty who phones up and pretends to know more than he does. It shouldn't be that way I know - all customers should be treated equally - but it's human nature.
The answer to your non-rhetorical, non-flame question ;) is that to a certain extent we do and to a certain extent we shouldn't have to.
We do notify the user in the sense that we give a meaningful error message. We have error messages that say things like "Couldn't write file X - check disc space and permissions" and we still get calls from people who haven't checked either.
Let's also distinguish between end-users and IT staff. The expectation is that I am dealing with IT staff not end-users. A high-end RDBMS is a complex technical product which requires some administration. So I would hope to be speaking to technically competent people. People who know how to interpret straight-forward error messages.
You see there is a learning curve to proper administration of an RDBMS. I think a lot of people think you can take that learning curve away by good interface design to your tools. You can't you can only reduce it. And in fact the more you hide the underlying complexity the more you discourage people learning the concepts they will one day need.
But another point about all this is where do you want all the error-prevention? A lot of the features and architecture in any decent RDBMS is about making sure your data is consistent. An awful lot of prevention happens behind the scenes that the user is never aware of. But that prevention is focussed primarily on safe-guarding the data - not necessarily on preventing all errors. So if your filesystem fills up the database processes might fail - when you restart them your database will be consistent. Trust me that's a major reason why people buy high-end RDBMS's.
What do you do instead?
Seriously - if it can't write to a filesystem it needs to write to what should the RDBMS do? There are 2 options - one is to stall all writes until someone spots the error message and takes action - the other is to exit completely. In the first the system would appear to 'hang' - in the second to 'crash'.
Generally speaking I think that the best design is to try to keep the system up for as long as possible - so you should 'hang' - but as soon as there's a risk to database integrity you have to shutdown. Sometimes you just can't guarantee the data integrity anymore (because of lack of resource) and you have to deliberately crash.
MySQL is fast, very fast, but the application becomes slower: .. it becomes a decentralised data integrity model.
No subqueries.
MySQL has a speed advantage, but that is because they didn't choose for data integrity options, in favor of speed. (think about lack of roll back buffers). Instead of using one complex query mysql users have to write serialised less complex queries and pass data on the application level. This drops the speed advantage on MySQL.
No stored procedures.
Again MySQL can't perform operations in a block, and queries have to be sent in a serialised manner
No triggers or foreign key constraints.
Data integrity has to be provided on application level. Just figure 15 developers
Only table-level locking.
Row level locking has to be simulated on the application level.
BUT they're working on it....
Oracle oracle oracle - checkout what the really mission critical systems run and you will find most are db/2 on os/390 parallel syplex machines. All the credit card companies use os/390 or as/400s where the rdms is integrated with the hardware and the os.
For web sites with small or medium complexity I realyl like using SQL Server. I'll probably get ripped apart, but things like stored procedures really made me like it a lot. Also, for beginners or for people who aren't DBA's full time the GUI tools for creating datagbase and tables really help a lot.
Still, with a plan, you only get the best you can imagine. I'd always hoped for something better than that. -CP
Im my career I've worked a lot with Access, Oracle and SQL Server. In our organisation, we've made a strategic decision to move away from SQL Server to embrace Oracle, as we wish to centralise and store our data (we publish legal information for corporate lawyers) which consists of large documents, plus financial, customer etc information. We chose Oracle (at a helluva price in the UK) simply because it allows us to embrace Java, XML etc and run it all on a SUN system which will give us the power to scale up as much as possible. We're just a small business (80 staff) but the sizable investment in Oracle allows us to protect our data and leverage it into the future and to protect the growth of the company. I don't like lining Larry Ellisons pockets like everyone else, but when it comes to the future of our business, at least I'll sleep at night.
Unless you are a M-A-J-O-R account (in purchases and/or in visibility), then having access to an Oracle-sized support service is the same as calling home to Mom for help [assuming it is a bug in the s/w].
When was the last time you heard of someone getting a problem fixed with the @Home network? For the most part, a fix there is usually just a server reboot or -HUP. Fixing code is a fair bit more difficult to get done.
Personally, I have to make this very descition several times a year for discriminating customers who only wants waht's best for them.
/. before... I just can't seem to find it in the Database :)
Problem being that 9 of 10 of our customers choose a UNIX platform, and as such MSSQL is out of the question.
Aditionally, they all want at least 16 hours of support every day (usually excluding Sundays), and some want 24/7. This leaves out the more or less free DB's.
This usually leave me with only three choices (four if it's an SAP R/3 system the customer is running, 'coz they have their own large scale DB only for SAP R/3):
Oracle
Informix
DB2
And since I hate administering an informix DB, I usually recommend one of the others (most frequently Oracle, as I can get better support for it). And in these cases YES they ARE worth the cost !
For smaller custumers on NT platforms or LINUX, we still most often go for Oracle or DB2, mainly because they're easier to convert if the customer decides to change platform from NT to UNIX or vice versa. MSSQL is just out of the question.
We have one customer on a NT platform with a 2 Terrabyte DB, their main problem is that NT cant handle anymore logical drives, and therefore the DB can't be extended ! If this had been on MSSQL, we would have a SERIOUS problem convering the data to a UNIX box with another DB, well, at least a very TIME-consuming problem.
But for the smaller customers, say my brother who runs the IT department of a small company, yeah, mySQL is fine for him. Only 15-20 concurrent users, load is low, data entry is minimal, support is well, not that needed, since there's very few problems with a DB You don't use a lot. For very small DBs You Don't really need the huge ass expenses, and as such You should not buy the "brand"-name DB's like Oracle. But it's all a matter of usage. Often You'll find that paying the bucks is actually something You need to do to be able to provide the level of security / support / availability / speed that the project requires. And most often, the big point of sale for these huge ass DB's is the scalability. Like I said earlier, You might run into a situation where You need to be able to scale up Your DB, without the loss of any of Your business requirements needing to be "altered" to conform to Your choice of RDBMS, instead of the other way around.
But this discussion has been on
--- To err is human... Am I more human than most ?
It's horses for courses. If you're running a web-site like Slashdot, MySQL/PostgreSQL should do. If, on the other hand, you're running a bank or telco, you need the big guys like DB2 or Oracle for consistency, reliability and speed. And forget MSSQL, it's not a true enterprise DB yet.
Try processing 100 million phone call records in a 4 hour batch window, or storing a few trillion dollars worth of bank accounts in MySQL and you'll soon see what I mean.
Hell, there are times when even the big databases can't cope, and we have to use flat files and assembler code. You choose the best solution for the job, whatever your personal preferences.
Seems Slashdot cannot handle critic.
JDBC is an API for using databases...it is not actually a RDBMS. If you use JDBC, there still has to be some "real" database behind it -- its not the Java equivalent of MySQL or MSSQL.
Linux is really just getting good now. I think the open databases will follow the Linux upward trend. Until then, Informix is a good choice for a "big time" database. The free ones will be good for most jobs though. Quote heard in the hallway at work - years ago "Oracle is the most overrated software I've ever seen."
PostgreSQL is probably more in line with Oracle in terms of functionality and performance. Last I saw, PostgreSQL still lacked a few SQL commands that were included in Oracle, but that these were in the works.
I've just gone through a similar situation at work. The thought was whether we could use MySQL or Postgres for our application, or go to a commercial grade product such as Oracle. DB2, Sybase weren't given more than a glancing consideration. We went with Oracle.
.com to a more substantial health care company.
I have worked on both fronts. I was a PHP/MySQL developer, and now I'm a (very green) oracle DBA. During this time I also moved from working at a
The difference between the databases really is night and day. And people need to put their preferences aside, MySQL is really just a better way to read and write data without actually writing it to a flat file. I mean its a really good way, and all you will need for a website. I remember thinking that I was all that and a bag of chips because I could store a picture inside of MySQL and pull it back out with PHP. So I did it several hundred times with thumbnails. All was well until a little bit of corruption got into the system, and they all got screwed up. No one ever mentions how easily (at least in my experience) mySQL can get corrupted. Even so that's all dandy for website pics, but what about data that really means something to your company? For websites, and with the possible exception of ecommerce sites, you may want to use mysql. For any mission critical system you _need_ to have an industrial grade database. Data is the life blood of your company. This is true in EVERY (Billing, Human Resources, Supply Chain, etc) conceivable way.
Only a fool, or someone who wants to prove an open-source point (or perhaps one in the same) would think otherwise.
PGSQL is a wonderful database, as is MySQL. They are wonderful solutions for some applications, but not mission critical ones. And that doesn't mean that you need to spend 2.3 million on enterprise edition of oracle on an E10000. However, DB2, Sybase, Informix and even the Standard edition of Oracle are fine examples of high grade databases. I personally would not run MSSQL because I believe you're defeating the purpose of providing a high grade database in a 24/7 environment (for god's sake run some sort of unix variant). And of course the feature set of the non-open source databases are more robust than anything open source. Triggers, functions, generally a 4th generation language for programming functions in the database (and stored procedures). The big guys are also embracing java in very big ways. Transactions are very important these days, and honestly these databases provide much better security features than any open source database that I'm aware of.
I'm not saying that open source databases are bad, I know they have their purpose, but on some levels they cannot compete with companies that throw millions and millions of dollars into database research, and I don't think they should. They should just do what they do, just really well.
Because you've never developed a large online system that uses a database. Try doing that before asking what the heck a large database is needed for.
Access. Sheesh.
Yeah, lets just go slap on a new OS. How often does that happen in production. Give me a break
it happens from time to time - and then you are *glad* when you can just export the database and import it into the new server. Where I worked, we had in the last 10-15 years the transitions VMS ->DOS,Novell -> WinNT -> Linux with Oracle DBs and there were no major problems. Sure you have to be careful with version transitions as there can be subtle changes that break one or the other application but you don't have to completely rebuild your DB from scratch.
There is something interesting at www.oracle.com/guarantee/.
Back in 1997-8, I did some work for the USPS. The systems I worked with at the time used DB2.
Perhps they have since ripped out all the DB2 systems and replaced them with Oracle. Heh... maybe this explains all the recent postage increases.
I take drugs seriously.
Yes, but will MS SQL server run on an Alpha? Will it run on OS/390? Will it run on a Sun E10000? Will it run on all major platforms, sharing data without interopt issues, with a sizeable stored procedure base?
No.
Microsoft SQL server is great for most light companies, or those that run homogenous NT/Alpha NT/x86. Here, we run S80, H80, 390, GS-320, and work directly with customers using SPARCs.
Here's a benchmark for you: http://www.oracle.com/ipo/index.html?fastest.html Unlike the MS Datacenter server (which is largely vapor ware) Oracle has been in use on major hardware for years. Oracle also allows you to choose which OS and platform work for you. MS SQL Server does not.
I'll give you middle-weight, on data center server, but until a Database has a signifigant installed base on the above $1M mark systems, it's not a heavy-weight. Perhaps MS should consider releasing a version of MS-SQL for other operating systems. MS-SQL could probably be a contender on VMS or 390.
-WSAn operating system should be like a light switch... simple, effective, easy to use, and designed for everyone.
The team I am on runs SAP for about 300-350 simul. users, and I would not trust anything but Oracle (64-bit) or DB2 (64-bit) to do the job. We have roughly 3 million SQL statements per day, with about 25,000 SAP Logical Units of Work.(reports, that is) The DB2 RS/6000 systems we have are just up to the task, and we are in the process of getting bigger guns. I say that if you need sub-second response time for many thousands of transactions, you need a real database. If all you are storing is your personal CD collection, or data that really doesn't require 8+ million column deep tables, and terrabytes of storage space, Why Bother? Postgres and MySQL, and even MS SQL are very good lightweight SQL servers. -WS
An operating system should be like a light switch... simple, effective, easy to use, and designed for everyone.
And if i can imagine NASA running MySQL, yeah, i can imagine Visa or Mastercard running it too. The sky ain't the limit :o)
This sig is intentionally left blank
You are right, but the DB product is great. Actually Oracle sales an excellent DB, but most of the other stuff sucks.
You guys need to understand that in real systems its likely to have gigs of data, thousands transaction/sec, thousands users and hundreds applications.... and plenty of communications, machine, DB, clients, etc... hang-ups, crashes, etc. trying to handle all that complexity at the application level without ACID properties, is just impossible. The real value of a system is DATA, other considerations as OS, machine, coding, language, etc.... is meaningless. And DATA integrity only can be 'secured' with ACID transactions. You can use a DB without ACID only in monouser, small systems, without critical info... in other works to play a bit.
Sorry, couldn't resist.
It sucks, but it's going to take more case studies like NASA with MySQL in order for free software solutions to not be laughed out of the corporate world. Accountability is just as big an issue as performance and scalability and the other standard RDBMS things.
When you've got millions of dollars worth of information and something goes wrong, somebody's going to want to be able to sue somebody. It's unavoidable. Also, you've got to wonder how readily an insurance company is going to support your free software adventures when, in the end, there's nobody that can be held accountable in case something goes wrong. In the end, this factor alone makes MS (let alone Oracle) seem a more attractive option than PostgreSQL, MySQL, etc.
--------
Bleah! Heh heh heh... BLEAH BLEAH!!! Ha ha ha ha...
I think that they have all kids of separate databases. Like any big company, it's tough to implement a company-wide standard. The group that I worked on used Oracle 8i on Sun ('98-'99)
The USPS has many different Oracle instances (it's the only database used in USPS), and many of them are larger than that.
Plus, DB2 can run on many, many, many platforms. This gives it extra portability that other RDBs can't match.
They stuck me in an institution, said it was the only solution, to...protect me from the enemy, myself
MYSQL can scale fine as long as you know how to scale Linux. Buy a few extra boxes with the money you save and distibute your load
"The Most Fun Possible on 4 wheels" is at SunBuggy in Las Vegas
In the past five years or so, I have worked a lot with RMS Flat Files (VMS), Foxpro, MS Access, SQL Server, Sybase, Oracle, mSQL, and PostgreSQL. I am not an RDBMS expert (that's Bill - he's worked here for 20 years) but I have formed a couple opinions.
If you've never worked with Oracle, you should step up to the plate. While most web site you'll design won't use half the bells and whistles, if you do they make all the difference in the world. SQL is a 4th Generation Language (4GL), but Oracle Enterprise Edition also comes with PL/SQL, a procedural 3rd Generation Language (3GL). Now you're not only asking the db for the data, but you're telling it how to do it. And let me tell you something - it's FAST! They also have ProC, ProJava, and ProCOBOL 3GL engines.
Oracle has amazing point in time recovery, which has saved my ass more than once. While Postgres (and even SQL Server) have some transactional support, Oracle gives you complete Point in Time recovery through Rollback (before image) and Archive Log (after image). This is a must if you're running a big eCommerce site or a mission critial database of any kind.
The big database support a lot more tuning as well. Oracle has two SQL Analyizers - the newer Cost Based and the older Rule Based. To keep the Cost Based Analyzer running smoothly, one needs to reanalyze the tables after significant changes in the data. Oracle supports additional types of indexes based on the distribution of data being indexed:
SELECT * FROM TABLE_PEOPLE WHERE SEX = 'M';
Well, if table_people represents all the people in the US, it's going to be around 50/50. If table_people contains members of the US Army, it's going to be over 90% Male. Based on those distributions you can decide what kind of index to use.
You also get transportable tablespaces, partitioned tablespaces (for those 1 billion row tables you have), packages, objects, and nested tables (tables within tables). If your database is in Oracle, if it's designed well, and you make proper use of all the features inside, you will see a huge difference. But that represents a large development effort, even for a small project.
But most people are designing a web site with a dozen or so database tables, and then what's the point? A couple stored procedures, a view, stick to mSQL or Postgres - they light and fast. They'll even give SQL Server and IIS/ASP a huge run for their money (SQL Server is really bloted, but still a lot better than it's grandfather, Sybase).
fooguy?
"All I ever wanted was to see Larry Wall give Bill Gates a Perl necklace."
http://www.eisenschmidt.org/jweisen
There are more than two options.
First, one could release the reserve storage one had allocated previously, guaranteeing that requests already in the pipeline could be serviced, while choking off new requests until the problem is resolved.
Second, one could send an automatic email notification to both the system administrator and to support, alerting the admin what the problem is and prepping support with information about the system in case a call comes in from the site.
Third, the system could trigger a purge, compress, or archive procedure which might be able to resolve the problem without operator intervention.
There are other ways to approach the problem as well, such as checking for disk full conditions before they happen and requesting preventative action. Rather than list all the possibilities, I just wanted to point out that there is a tunnel vision in traditional engineering approaches to error handling.
Part of the reason for the tunnel vision is that advanced error handling is just not a status job. No one gets as excited about doing something like this as they do about coding a new feature. But there are many more things that systems could do for the customer than they do now, and many of these things have a serious value proposition.
Tim
Still, why should they have to do the checking? Those are both software-testable conditions. This is an example of error messages by guesswork, one of my pet peeves. Such messages are often misleading, because the problem may have occurred for none of the reasons listed; and these messages are always lazy, because it would only have taken a few minutes to write code that checked to see if the problem really was that the disk was full or that there was a permission problem.
(In fact, the error message code often ignores the actual error return from the routine that failed, and even though there was, say, an out of memory error code, it still blithely goes ahead and says to check permissions and disk space.)
the more you hide the underlying complexity the more you discourage people learning the concepts they will one day need.
Sorry, I really don't think that putting out accurate error messages hurts anyone in any way.
The error message concept itself can be an example of tunnel vision. On a server system, no one is sitting in front of the console and no one will see the error message. Instead, errors on unattended systems need to be routed into email or instant messaging, so the responsible person will see them wherever they are.
You do have a good point that data consistency algorithms like two-step commit are themselves a form of error prevention. I'm just saying that there is still a long way to go in error prevention, reporting, and correction, and that the obstacles to moving forward on those fronts are cultural and psychological rather than technical.
Tim
This is not meant as a rhetorical question, and even less as a flame. I've just noticed, as a longtime user experience designer, that many support people's complaints about user error never need have occurred in the first place, and could have been avoided by the well-known user experience principle of error prevention.
Regards,
Tim
(1) Development effort has benefits that are replicated on every user's system, so the most economical place to address problems is during development rather then by deferring costs onto the users.
(2) Error resolution and correction are inherently far more expensive than error prevention. Human society has known this for a very long time, hence the cliche: "An ounce of prevention is worth a pound of cure."
As for inability to deal with every possible error case, I generally agree with you, but in this case we are talking about an extremely common event, a storage device becoming full. Most complaints I hear from support staffers revolve around problems just this common, which the system could have guarded against but did not.
I think there is a psychological issue here, in that there is an actual pleasure -- or at least reward -- in blaming users for their their problems. There's a feeling of superiority in it. That (and related blame and dominance issues) are among the reasons that error prevention has not become a more widely applied principle.
Tim
SAP DB is derived from ADABAS/D 6.1, which ran on Linux and many other platforms, and is a RDBMS. ADABAS/C, the MVS database, is a non-relational product and entirely different.
is like comparing a moped to an 18 wheeler. Sure, you can do delivery jobs with either. But how much do you need to deliver? And how fast?
-
Actually AC, yes it is. Relational DataBase Management System From: http://www.mysql.com/information/index.html MySQL is a relational database management system. A relational database stores data in separate tables rather than putting all the data in one big storeroom. This adds speed and flexibility. The tables are linked by defined relations making it possible to combine data from several tables on request. The SQL part of MySQL stands for "Structured Query Language" - the most common standardized language used to access databases. Now, to the topic at hand: Is mySQL a far cry from stable oracle releases (read: not 11i)?, yes it is.
According to Dr. I. Snivvel of Carnegie Mellon University, the selection of a database, and any software for that matter, should be based upon the CRUD (Conspiracies, Rich Republicans, Ulcers, Da Money) criteria.
Conspiracies: Needs no explanation. Just look at Microsoft and Sun. It has, in fact, been proven that the Windows NT/2000 kernels are built around PROMIS software.
Rich Republicans: Simply take a look at the software to see how many rich people it has helped create. A significant number of rich people can indicate that the software is widely adopted by business and government, which is usually de facto proof that the software must be crappy. A piece of software is only as smart as the dumbest person capable of using it.
Ulcers: How close will you come to a mental disorder or suicide while trying implement, customize, or develop with the software? Software that pretends to be easy to install and implement often is, but can be insidiously difficult to manipulate to the users desires. Take a look at the documentation for the software package -- Is it full of warm and fuzzy screen shots that explain how to right-click the mouse? Or is loaded with LISP references? Is LISP mentioned at all?
Da Money: This subject requires an alert wariness on the part of the individual acquiring the software. Software vendors will often try to obfuscate the crappiness of their software by charging large buckets of money for it. Their marketing personnel often exploit the behavioral principles revealed by North Korean researchers shortly after World War II: That people seem to think that they get something better because they pay more for it. The purchasers will readily adopt 'Sweet Lemon' rationalizations to explain their purchases. On the most occult purchasers realize that something free is always 100% cheaper than something you pay for.
The only thing that we learn from history is that nobody learns anything from history.
US $20,000/processor for enterprise edition. Maybe you meant Access by itself.
As for why? This is very simple. We are risking the integrity of the data that keeps a company alive. We have to be very conservative when putting all that data to risk. Personally I will only recommend Oracle 8+ or SQL Server 7+. I started playing with Informix Redbrick warehouses about 6 months ago and true, it's expensive as hell but (1) it's bulletproof and (2) our client had it already, so it did not impact the bottom line of the contract.
I will consider posgres as long as I have trained personnel available to deal with the server (we are a Microsoft-only shop for now). I will not consider mySQL until I get solid proof that it passes ACID and it beats posgres.
To be honest, I won't even recommend Oracle anymore. If the client has a license for Oracle already paid for and gathering dust we'll use it to our advantage. If they have to buy licenses we'll recommend SQL server every time.
Pedro
Pedro
----
The Insomniac Coder
So for all you Oracle hackers (I know absolutely --zero-- about all databases, having never used one in my life on the business end), when you call support at 2:30am make sure the lady helping you says "hi" to OYAAAAAH
"Champagne for my real friends - and real pain for my sham friends!" http://ericblade.postalboard.com/
About 6 months ago, at my local LUG, a company called merant (www.merant.com) talked about releasing their database program sometime soon (october) under the GPL. They said they'd really like support from users like us to do it and they'd survive as a company by doing support. Does anyone know if they've actually done that or not? If so, why not switch to it for mission critical apps?
-- Political fascism requires a Fuhrer.
These prices are at least a year old. I think your certified reseller was shonking you.
I said (and I quote) "Oracle Standard Edition on that hardware is a few thousand dollars a year". If anything, I overstated it, as a 2-year license is $3k.
"Patriotism is your conviction that this country is superior to all other countries because you were born in it." -- GBS
Is this some sort of a joke? It certainly has to be. Why would Bush be reading, let alone posting, to Slashdot? :) Sheesh... had me going for a moment there! ...yeah, down with COMMIE err uh, Commercial software...
First off, MySQL, SQL Server, and the like cannot run large data warehouse-like stuff. Think parallelism, striping, hot failover, and the like. As well, REALLY large transaction-processing systems (banking, airlines, etc.) need the Oracles and DB2s of the the world.
As for websites, I think these large indistrial-grade DBMSs have a place too. Consider a Web app server like iPlanet or Websphere. To support transaction-intensive environments (again...banking, airlines, etc.), the session management requirements are HUGE. I don't think I'd trust MySQL for something like that.
Am I right or am I just ignorant?
CrazyLegs
"Pork!!" said the Fish, and we all laughed.
You can't lump Sybase (or rather Adaptive Server Enterprise as it's now known) into the same bag as MySQL.
Sybase is a high end RDBMS and a major competitor of Oracle. Although Oracle has a larger market share overall, ASE is the leading database in many industries, including the financial sector in the UK (over 70% in fact).
And it costs a similar amount to buy, but you get what you pay for.
Pardon my ignorance, but what's an SEV 2 TAR? A runtime error message?
As for the "user error" response - it does happen I'm afraid but let me say this. I get so many calls from people who are supposedly IT professionals who have no clue at all that it's real hard to stay un-cynical - but I try.
Um, even IT professionals sometimes have to handle an system "on the side" they don't know much about.
Has happened to me, too: I'm writing code for a very small company where I'm also the network and database administrator. While I know NT reasonably well, I didn't have the time to properly learn Oracle administration - which promptly led to a somewhat embarassing support call.
Thanks to the guy at Oracle support, however, we got it fixed nicely. He sounded a bit frustrated, but quickly found the problem.
So if you get another call like this, don't get mad - these things happen.
C - the footgun of programming languages
I guess speed is an issue, but the fact is, things like Oracle supposedly *can't* run on my computer, because I don't have enough memory... But if you're going to be running a database on a Cray, you can probably afford to hire a small committee to invesigate different databases. ;)
________________________________________________
________________________________________________
suwain_2
It really all depends on what you want to do.
:)
If you are making a small site to target a small community (say, the BeOS community), then a small DB such as MySQL is more than addaquate. If you lose some data here, or the database server goes down, it's not a major loss. So you lose some banner impressions, big deal.
If you run a higher traffic site, you might use Postgres.
If you need scalability and some more advanced enterprise features, you then use Sybase or MS SQL (not sure where DB2 falls in here, as I don't have much expereince with it).
For mission critical, can't ever be down, systems, you should *seriously* consider Oracle. This week, I am in a Oracle DBA/Architecture class and the instructor has a few DBA friends at First Union (an American bank - a rather large bank at that). She said that their online banking/db system is run by *57* servers using Oracle. Why? Because just *one* hour of downtime costs $10million USD. Yes, TEN MILLION. Would you use Postgres in that situation? I don't hardly think so.
As well, if you need to be able to recover *everything* from a crash (disk/db/system, etc) then I would definately recommend considering Oracle. You can easily backup from the latest backup and use the transaction/redo logs to bring the database up to data of the very last transaction before the crash (you might be able to do this with other RDBMS as well - not claiming to be a guru by any means).
Again, it all depends on what you want to do. I certainly wouldn't recommend Oracle for running a small time site. I honestly don't think it would really even benefit Slashdot to use it (for the cost).
just my opinion, feel free to flame
Most recently, I was providing support for a major fast food franchise. Many of the managers did not know the difference between a modem and a computer (and one claimed not to know what a keyboard is). They have $100k-worth of networked computer equipment, and they are letting people run the store who don't know what it's called! Not all of them are that hopeles; half the managers know at least some basic concepts.
I don't mind getting calls from new managers who are a little rattled that the printer isn't printing (because the power switch is OFF). What bothers me are the people whose systems are messed up due to bad maintenance, or those who expect me to give them a one-sentence answer that will have everything repaired in a few minutes (especially those who expect the answer without doing any troubleshooting), or those who expect me to be able to fix any problem by dialing into their store. Then, there are those who expect free work (a LOT of free work, as it turns out--business-model significant amounts), or expect the help desk to do their day-to-day menu items, or expect a staff of 30 people to give 8000 stores immediate priority when they call, or expect a technician to arrive onsite within minutes of calling the help desk.
Some guys at Avid (the video editing software company) produced a funny short about customer support that speaks across industry boundries. I saw it at I-film; maybe it's available at other places.
Taking stuff apart since 1969 (TM)
I saw a couple of comments lambasting the poster for saying the word "Access". I am sure that he only meant using Access as a front end, not as the acutal database itself.
As I write this, there is a banner ad for safeWeb at the top of my screen. It depicts a person's eye peering through a hole, and a finger comes along and pokes the eye. Now my eyes hurt from just looking at it.
When I was investigating how one backs up interbase, mySQL and PostgreSQL I found that only interbase has a backup utility that can be used without taking the server down.
Yes postgreSQL has pgdump, but if you read the documentation it has a few holes in the implementation that you need to be aware of.
So if you want to minimize downtime I would use interbase or one of the commercial alternatives.
Just my $0.02
Who's paying? Will you get fired if an Oracle DB breaks? Will you get fired if a MySQL DB breaks? Seems to be a no brainer.
I think the only upside of the proprietary database software is that it has a known development team and so it's a more 'stable' and somewhat 'safe' development. This said, i think free dbase software blows the others away. It's free, runs on multiple platforms, comes with your Linux, you can download it anywhere, it's not big and all. And it was a problem that got solved in Linux when MySQL for instance came up, there was finally a strong dbase available for us.
The value is in the data, not in the hardware and software that maintains it. Once your data is worth a few billion, you cannot afford to take even ONE chance that your database system isn't up to snuff. It totally depends on the value of your data and your need to protect that value.
I found the following on a MySQL site.
http://www.mysql.com/information/benchmarks.html
In particular, look at the differences between Oracle 8 and MySQL about 2/3 down the page. I must say that I am not a database expert, nor do I suggest that anyone make a decision based on one source, but the numbers are interesting nonetheless.
Just keep it simple.
I think for a small mom'n'pop system you could get away with a flat file & in-place record edits!
_ db_linux.htm
Oracle/SQLServer come into their own when managing hundreds/thousands of concurrent users hitting terabytes of data - you need:
1/ High quality query optimizer (to efficiently select the best access paths (data+index data) to satify the query - mySQL sucks at this for complex queries; Postgres is improving
2/ Online dba tools, you want to backup the db without impacting the online load; reorg indexes without stopping access etc...
3/ Instrumentation, I want to know what is going on in there, so SQL traces/perfmon or ps info is essential
4/ Security - prefereably hooked into the o/s so I get all those nice C2 features like password attack lockout, easy auditing and granular control of data access (by col/row/even data value) etc
5/ Availability - if the db isn't up my Customers ain't making money - don't want a cpu crash to take me out..
6/ Effective use of serious amounts of cache (32GB)
7/ Support, support, support
When you've got Boeing's complete inventory, or Bank of America's balances, or American's seat allocations you can't lose data, ever - fraid I don't trust mySQL that much yet.
However I think the SAP/DB opensource project (based on the Unix version of Adabas (an MVS database that rocks!)) is looking promising, SAP will even sell support.
See http://www.sap.com/solutions/technology/sapdb/sap
**Vanuatu or bust**
Bah Humbug, there's plenty support for free software. The difference is that at this point commercial solutions such as Oracle are still technically superior.
Both MySQL and PostgreSQL loose in areas that Oracle takes care of.
Postgres is ACID compliant, but it's slow, especially in a situation that requires a lot of new connections all the time, such as a high traffic web site.
MySQL is faster, but lacks a lot of the proffesional features present in Postgres or Oracle and is not ACID compliant.
At this point both free databases are good and very useable for a lot of tasks but for some especially large things Oracle is more suitable.
For now......
What evidence do you have that NASA uses MySQL? Even if NASA does make use of MySQL, people tend to be impressed by anything if "NASA" uses it. Do you honestly think that NASA has to deal with databases as large as Visa? Is NASAs use of something really taht impressive? Earlier there was a /. article detailing how the international space station may be using laptops for life support.
Should Visa run their MySQL database on laptops?
Furthermore, do you think that NASA exclusively uses MySQL for ALL of their database needs? It is a matter of using the right tool for the right job.
Computer Science is no more about computers than astronomy is about telescopes. --E. W. Dijkstra
Nobody ever said Ellison didn't know how to catch the headlines...
here's one of many links that nails it to the wall...
I manage a small group of Oracle developers. I hate Oracle. It's dog slow relative to MySQL. It's the most complicated piece of software I've ever dealt with. When you discover bugs, you're in for pain. And it'll suck you dry of cash faster than heroin.
So why do we use Oracle? Because all the others are toys. Without things like:
Scalability
Recoverability
Serious support
Hot backups
Replication
Basic features like transactions...
And so on, ad nauseum
If you require these things, don't go trying MySQL. MS SQL has much of this, but it's an unreliable POS, and what it doesn't have makes it unusable for serious, industrial strength use.
MySQL has its place, of course. It's kickass fast, but what DB wouldn't be without all the overhead required for features that would actually make it heavy-duty? We use it for our website; mostly read-only stuff, for which it's highly suited. But without support for things as fundamental as transactional atomicity, it and its ilk are worthless for the big time.
Oh, and MS Access? That's kinda like trying to fly an Estes rocket to the moon. You would have to be a moron to compare Oracle and Access seriously. You're talking watermelons and grapes.
Oracle's license clearly states you have no recourse even if their software destroys your 3Tb database.
There is debate wrt whether the cheaper/free products can scale up to the size of Oracle databases, but most of it is more than likely FUD. Oracle refuse to allow proper testing and independent testing that has been done without their permission has been inconclusive; that tells me Oracle has something to hide.
OTOH most people will cover themselves by buying Oracle; at least that way you can blame them, afterall it is a well known, expensive product.
You do get a lot of what you pay for with Oracle on both ends. As a developer, after working with Oracle exclusively for many years, I have certain 'expectations' out of a database. MySQL and PostgreSQL are both excellent databases and I will likely go with MySQL as a starting point to my own business, but once the $$$ can support it, I will likely upgrade to Oracle because by then my business requirements will be significant in my RDBMS back-end. I will require the scalability, reliability, and support available from going the high-cost route with Oracle.
I found some rather interesting numbers when looking at MySQL vs PostgreSQL, esp. around the total number of concurrent connections both can handle before dying. From this chart, there was a certain point at which MySQL core-dumped and died. This may be different now, but the article did reveal that MySQL while a great GPL DB engine, is not qualified for a true "enterprise-grade" database solution.
My advice would be go with what you can afford, like MySQL, and just practice the proper application development techniques to decouple your app as much as possible from the db on the back-end.
My top 3 wish-list items for MySQL would be sequences (not auto-increment columns.. i hate those), stored procedures and stored functions.
-Azzy
--
$ chown -R us:us yourbase
In my experience it has been silly to sepearte a web interface to your datastore and a traditional application interface. It makes no sense. Sure, so you have a 3TB db, why would you need to have a different db if you wanted to add a web interface for your 'cute' website? The whole point is that the datastore never changes, but the interfaces grow and are added and get more and more powerful. But more to the point if you still believe that websites are cute and not mission critical, what planet have you been hidding on? Oracle does 70% of its trade online. I would say that 70% of your revenue stream is pretty mission critical. What about easyjet.com. They do not exist except on the net. Amazon.com. Yes, their website is just a corporate self promotion gimmick. Not mission critical at all. Go back to your donut, monkeyboy.
postgres works fine for me.
I would *never* think about being locked into a proprietary solution; Do you really take me as an idiot?
I hate to say it, but I haven't seen anything under *nix that will do CASE work.
I have experience with only two CASE tools, so my input will be limited, and both work strictly under M$ Windows. I've worked with Popkin's System Architect, and in my experience it is the most bug-ridden PITA program I've ever used, in any category.
I'm currently using Oracle Designer, and I've been highly impressed, with a few reservations. Designer allows you to draw your ERD, then run a transformer to generate the physical design model. After you run the transformer, you can put in all your code, indexes, triggers, etc.
Then, you press a button, and your DDL scripts come out. Dead easy.
Theoretically, it'll also let you reverse engineer from a connection to a database. I say theoretically because I've not tried it.
Designer (obviously) will work REALLY WELL with Oracle. It'll allow you to generate Oracle-specific code (for versions 6 and up, IIRC) from, basically, soup-to-nuts of your back-end: table creation, constraints, indexes, triggers, and stored procedures -- in both PL/SQL and (theoretically, I haven't attempted it yet) Java.
Designer will also output SQL dialect for other varieties of databases: SQL Server, Sybase, Informix, DB2 are all included, IIRC. But most importantly for you, it will also export in ANSI92 SQL. So, PostgreSQL (I don't have any experience with mySQL) at least will be able to use the generated scripts with (hopefully) minimal by-hand modification.
I have some reservations regarding Designer, in that it seems to have several bugs. I say "seems", because it looks like there may be workarounds (or "the proper way to do things") that I simply haven't figured out yet. I've not found them to be deal-breakers as a tool choice yet, mostly because I can (and do) always go in and edit the resultant scripts by hand to tweak things here and there. However, my volume isn't high, and I don't have an unreasonable time commitment for completion. If you need your generated code to be 100% perfect all the time, well, Designer doesn't do that for me. YMMV.
Good luck!
--
My ID is in the 200k's.
n/t
"It must be something truly enormous, Trismegistus"
Large systems? What bout NASA? They switched from Oracle to MySQL and it seems to be working just fine...
Well IMHO I would go with Oracle on a Sun machine before I'd spend money configuring an i386 based machine to run SQL. Some of the issues that should come to mind are simple, cost, effectiveness, easy of use, scalability, etc.
Costs:
Oracle programmers almost always earn more than other DBA's but Oracle is a bit more stabler and easier to maintain than the SQL's. Sun machines are more expensive than i386's but some of the neccessary upgrades to get a stable machine would likely equalize those costs. Oracle offers great support, while SQL offers mailing list questions you have to hope get responded to.
Scalability:
Sun clusters using Oracle are good for scaling when thinking about 100's of thousands into the millions. At a company I worked for recently we had about 4 million users on an SQL based DB which would choke, distort tables, you name it. At yet another company I worked for, we had a cluster of Sun e450's running Oracle without issues.
Ease of use:
Well I've touched on MS's products lightly and I was not impressed and was quite pissed off at their "issues", which are too many to name. Being familiar with Nix based systems seem to fit my criteria, and although I am not a DBA, I have found troubleshooting Oracle based DB's easier than SQL based DB's, SELECT FROM FOO WHERE NAME LIKE HUH?
my two sek's
AO's DB of chicks
Thanks for saving me the trouble of looking you up to copy your info. TEE-HEE!!
Cheers,
The Inquisition
NOBODY expects the Inquisition!
No doubt that the parent was a trollish comment, but it's also one of the best nutshell explanations for why people pay extra for the Oracles of the world. Sorry to see that it was modded down so quickly.
1. Oracle will support you. No matter what time of day it is, Oracle tech support mans the phones and consistantly is there to make things work again. That's of primary import for a company with a nearly unlimited checkbook.
2. Oracle is currently a little slower, but has a significantly larger feature set (most importantly, multiple versioning and transactions).
3. Wall Street like to hear you say "our Oracle order database allows us to blah blah blah..." Don't underestimate the power of the banker! There's a really good comparison at this site
I'm the best IRC client ever.
I don't think any of the current Open Source DBs can handle much past 100gig and still query will any performance. MySQL may also be a bad example since it lacks several features to make it a true RDBMS. PostgreSQL would have been a better choice. For anything less than 100gig, Postgresql rocks.
I can't imagine trying to put a 1 TG database on anything but DB2 or Oracle. PostgreSQL is the closest to these on the Open Source side, but it has known issues out past 50gig of data. If we put a cap on the size of the database I would totally agree that there is no reason to use a commerical database.
For most web sites, Oracle is an overkill. I run mhy site, geeks4free.com, using MySQL, Apache and PHP on the BSD platform and it just rocks. I see commercial web sites that are slower than my site and they are spending millions of dollars for that!! If more dot-coms were using my approach, they would still be in business.
* * * Free programmers? Why not? http://www.Geeks4Free.com * * *
In all this discussion between insanely priced commercial databases and free open-source databases, where does Interbase fit in? Interbase used to be a commercial database, but is free now. Why is there no rush to get Interbase?
Galactic Geek- the guy who made Geeks4Free.com famous.
* * * Free programmers? Why not? http://www.Geeks4Free.com * * *
Oracle has fast taken a large chunk of the RDBMS market.
But in the recent years, Oracle has been plagued with Microsoft woes.
And the customers have been plagued with empty promises of sales people which the db cannot provide. I still have about 3 TARs that are there in oracle for about 1 year now, no resolution!
While Oracle is ok for the casual user, it does not really pass its own ACID-test. Each time they come out with 'NEW' technology, promised features dont work. And of course if it does not work, then its a feature not a bug. To cite an example, within oracle OPS option, the update from one instance to the other does not happen instantaneously. When the sales person was asked to clarify, his instantaneously was about 4-5 seconds. Unless the dictionary has been rewritten since last, I thought instantaneously ment 'snap' you got it. We needed it to work that way, but oracle 'features' dont allow it. Oracle database's definition of instantaneously is 3 seconds, the amount of time it takes to populate from one running instanc to the other.
Another thing is the maintenance that you pay for. Oracle has forgotten what customer support is all about. When I pay 25% to 50% of your oracle cost each year as maintenance, I would expect them to have a more friendly attitude, coupled with a desire to 'really' help fix problems that customers are facing. But all you get is pathetic bureaucracy. And the replies will make any decent DBA jump out of window. Thankfully my office is in the basement, therefore no windows!
One of the worst part of purchasing oracle is that you cannot access their bug database unless you are on maintenance. Duh! They are your bugs, so why should I pay for getting access to documentation about them?
Well, comparing MySQL with Oracle or a enterprise level RDBMS is like comparing a motorcycle with a jumbo jet. Both will most likely get you to your destination, and while MySQL is much lower maintenance, the jet has much greater room to handle future issues, while being able to expand to other destinations easily. It should also be noted that MySQL is not a relational database (RDBMS), unless the application logic behind the data enforces the constraints. This is not to say that MySQL is not up to many tasks, but its not ideal for applications like data warehousing and complex relational database models. I certainly would not think that the NYSE uses MySQL for holding its trading records.
Microsoft has already abandoned TDS in SQL2K, or at least Sybase complient TDS. In SQL 7.0 you had to have MSSQL 7.0 SP2 in order for Sybase to talk to SQL Server. Microsoft announced, with the release of SP2, that future releases of SQL Server would not support that functionality
That's just it, Oracle has made it's name by staking it's reputation on data integrity and rock solid redo logs, locking and supporting utilities. What this means is that they have stood up in court numerous times, "If the Oracle database is setup and properly configured, YOU WILL NOT LOSE ANY DATA." Surely the argument does get fine-grained, i.e. "you won't lose any data up to the milisecond of the catastrophe, etc" but basically the guarantee is of NASA proportions, and that is why Oracle has the lion's share of banks, governments and serious businesses. I don't recall ANY other vendor making such promises. If you know of any, please inform me; it's about time Software makers be held to the same industry standards as everybody else. --Of course Oracle has the worst command SQL interface, crappy extensions (Forms bullshit, web DB etc) but KaChing is more than a city... Cooker
I work for a high-tech company based in Little Rock, Arkansas. We have in our data center what is what is most likely the largest database in the world. We also host many large companies' core databases. Most of which are Oracle. From my perspective, it's an issue of reliability and support. (and the security of having somebody to point a finger at if things don't work right) It's true that open-source developed database projects are getting better and better, and in most cases, they would work fine for small and medium-sized companies. The catch is that the company has to hire somebody (usually a lot of people) to develop, setup, and maintain these platforms. While this is also somewhat true even for a large RDBMS like Oracle, Oracle is still responsible for making the thing work. Furthermore, if I am the CEO of company "X", and I am going to spend hundreds of thousands, maybe even millions of dollars on a machine to hold the database, you had better believe that I want the OS on that machine to come with a guarantee and support, not to mention the database software. As for Microsoft solutions. They suck. You don't have to spend much time on Microsoft's website before you get a big fat sql database error, or .asp errors. You would think that THEY would be the first to make sure their own stuff worked right, but they don't and they never will.
Also, good as open-source options may be, for large-scale operations you can't beat Oracle. When you begin to look at a large company's database needs today, you have to look at a database that can handle millions upon millions of records. I would surely like to know if ANYBODY in the world has implemented a several million+ record database with mysql on Linux. Don't get me wrong, I'm a big open-source supporter, but at the company I work for, they call a 50 Million record database a small one. The larger databases that Acxiom has developed contain hundreds of thousands of records, and can be on the order of SEVERAL PEDA-bytes! That's beyond huge!
Oracle, Sybase, Informix and even MS are recognizing the fact that the quality and scalability of freeware databases are increasing. To keep the gap between these two camps wide as possible these corporations are adding features and dropping prices like crazy. Its great to see. Because of the freeware DBs we are seeing a shake out of commercial DBs. Oracle for example has proven itself to have features and scalability that no freeware DB has yet. This same shake out is what we saw with web servers and web hosting. Companys like Verio (http://www.verio.com) are offering shared Oracle hosting allowing you to pay for only a subset of a oracle db without all the upfront costs or pains of Oracle SA and DBA work. I think we will see this with more databases and more service providers as database hosting becomes more of a commodity item. This is a goodway however to get your foot in the door if you need the power of a db like Oracle without all the costs.
You pay for PERFORMANCE, STABILITY, FEATURES...
MySQL, Msql, Sybase, MS Access and other similar
low end DBMS's can handle no more than 1M records
and 50 hits/second. Also you must shutdown your
database for hours each day to backup your data
so you can recover from the inevitable crashes.
Oracle and MS SQL Server can do a little better
and handle 100-300 hits per second, 100M records,
and run 24/7 with online backups.
If you want to handle really large databases you
need to use a system that was designed from the
start to handle huge databases, billions of records
and thousands of concurrent transactions.
The fastest database by a factor of 10 is a little
known system called Model204. A single system with
a single Model204 can handle:
30000 logged in users
5000 concurrent transactions
5000 transactions/second
2500 active databases/system
10000TB of data
4 Billion records/database
350 Million Hits/day
ORACLE database tables can be accessed by SQL commands which are embedded in standard programming language programs.
Definitely not unique. PostgreSQL, for example, has hooks in Tcl, Perl, Python, C, C++.
They aren't talking about programming hooks. They are talking about Pro* (Pro*C/C++, Pro*COBOL, Pro*Fortran) which allows you to directly embed SQL into an application without hooks. You basically run a precompiler over your code which validates the SQL itself, allows runtime diagnostics at the client level, clean type conversion, etc. Personally I prefer OCI, but that's just me.
Features which I believe Oracle has most of the open source competition beat in is clustering (run oracle on x machines and have it load balance), parallelization of queries across CPUs in your cluster (very nice for long running queries), resource management (you can allocate x bytes for y task if you want), the ability to index based on function (index upper(col) for instance) and runtime configuration & status (the v$ tables are scary they are so complete).
Now, Oracle has several drawbacks. sqlplus has no readline support (quite annoying). Certain types of simple queries are not optimized (select count(*) sucks). The OCI libraries are big and there is no small client-only installation (we built our own tiny library-only install).
The world is neither black nor white nor good nor evil, only many shades of CowboyNeal.
First off, MS Access is not a true multi-user database as anybody who has tried to use it as such as probably discovered. It also has mega-trouble scaling about 1 GB. I'm not familiar with free software databases so I can't comment.
But what are the key selling points of Oracle?
1. Highly scalable architecture. Terabytes of disk, hundreds of transactions per second. Oracle can do it.
2. Clustering. Oracle Parallel Server provides true failover capability.
3. Replication. Another data distribution/failover/backup approach.
4. Tools.
5. Large # of apps designed to run with Oracle.
6. "Value added" add-ons such as Oracle Forms, the geospatial option, etc.
7. All out Java integration
8. Interfaces to most other databases, support for remote queries and dblinks.
9. Ability to back up database while it is run.
10. Very portable.
11. General lack of bogosities compared to the competition.
12. Market share --> confidence in them being there for the future.
13. Support.
The knocks on Oracle have always been its high price - try buying Oracle for even a basic Sun E-450. It can run upwards of $300,000 - and the relatively DBA intensive nature of it.
For anyone who plans to develop highly scalable systems, I highly recommend studying the Oracle architecture. It is a textbook on how to do large scale computing. And for those used to the traditional Unix client/server computing paradigm, it can be quite an eye-opener.
I know that money isn't everything, but...
Oracle admins cost a bundle for a *reason*. They are well trained, know their software inside and out and if you watched one work on an Oracle system, it's amazing (like watching a greybeard work on a VMS machine).
And to plug oracle even more, you're not at the mercy of the OS -- Get me a copy of MSSQL for Unix and we'll talk, but if the NT server goes down, and you can't take it anymore, you have all sorts of other OS's to slap oracle on. This is true for most other RDBMS's as well.
However, it should really be noted that 90% of 'relational databases' are so unrelated it's really sick. We have actual relations at our work, and it all runs on MySQL, which is a pile of crap, IMHO.
No subselects, transactions, real locking. About the only thing that mysql does well is serialization and select statements, both of which I could do with flat file databases and a perl script, and probably get decent -enough- performance to be usable in an environment that uses mysql.
To put it simply, if you know what you're doing and have any experience with this stuff, go oracle or perhaps PostgreSQL, as the 7.x series is probably the best free database that you can get your hands on right now.
Stay FAR AWAY from Informix. I yet to see a slower RDBMS... Granted, it's stable... And if you work on lucent switches, you're probably going to have to deal with it anyways.
If you just need to store crap to belch out later on a webpage, sure, use mysql.
I broke Oracle8i on Solaris 7 (patched) and Veritas LVM 3.0.x this past summer on a Netra t1405. I can dig up the script that provoked it, maybe, if I'm lucky. I still trust Oracle pretty well, though. It may have been a Solaris issue, but that would surprise me. I've only heard of one other coredump with Oracle 8.x.x on a primary development platform (Solaris and maybe HP-UX?).
I didn't actually break Informix; one of my friends working as a DBA for Cornell broke that in production, was corroborated, and that's good enough for me. That was on AIX. (Are they still in business? Their webpage is up but that's not really a perfect indicator of solvency.)
Sybase was on Linux. I'll admit, I cheated a bit there. I'd be happy to try and break DB2 UDB on something like an S/390 G6 but I don't have one of those lying around...
Remember that what's inside of you doesn't matter because nobody can see it.
The EMC was a red herring, after a fashion. It's more for the purpose of "see this pile of money? that pays for the DISKS to run a real OLTP DB.".
I mean, this is Slashdot, not a technical forum.
I'd rather spec a Sun StorEdge/T3 array (for Oracle) or IBM's competitor to it (for DB2 UDB, presumably running on AIX) for performance, since EMCs are legendary for being butt slow once you blow through the cache. There was an article on adaptive caching RAID arrays in Oracle Internals awhile ago; someone built a hardware solution...
The Cheetahs now come in an X15 (15000 RPM) model, if you're rolling your own storage (ittybitty Oracle setups, PostgreSQL, or MySQL).
They're pretty nice IMHO, I have a few of them in RAID 1+0 here. Snappy. It's also nice to rack them in a 4U so you have room to put two Adaptec UW-160 cards in there and keep each side of your mirrors on a separate channel. 72GB mirrored at 15000 RPM... yeah, it beats up the EMC "solution" for that scale of a project. Obviously I'd start reaching for Sun Storedge JBODs after that.
But yeah, you're right, that was a spurious point.
Remember that what's inside of you doesn't matter because nobody can see it.
Oracle is definitely more scalable, etc, but I'd like to attack it from the standpoint of a developer:
:) Sure, the above query can be done in mysql, but it would take me more time to write, and possibly mean splitting it up in to two queries. I can't use mysql now that I've gone to oracle.
There are some big performance features (statement caching, stored procedures, materialized views, etc) that definitely help with large web sites, and there are others that help me get my work done, like subqueries. I haven't used MySQL in about 6 months, and up until then there was no such thing as a subquery, which can be handy at times. Take the following query:
SELECT
a.message_id,
a.message_text,
b.message_count
FROM
messages a,
(select message_id, count(*) from message_hits group by message_id) b
WHERE
a.message_id = b.message_id (+)
Oh, and that reminds me- outer joins too.
I also have been tinkering with PostgreSQL since it was Postgres95, and I couldn't agree more. If you haven't checked out PostgreSQL for a while you owe it to yourself to download the newest 7.1 betas and try it out. PostgreSQL has gotten a lot faster. In fact, unless your application is extremely select heavy PostgreSQL will almost certainly give MySQL a run for it's money.
It also has a great many of the nifty features that make the commercial databases so expensive, and the mailing lists are amazingly helpful. It certainly doesn't hurt to give it a try.
Oracle is an SQL database....
For every problem, there is at least one solution that is simple, neat, and wrong.
Reliability and scalability.
It is pointless to even compare something like Access or MySQL to something like Oracle or DB2. If you are even asking this question, it means that the above two words are foreign to you.
___
___
If you think big enough, you'll never have to do it.
Also (at least to me) is replication. AFAIK, there is no compare on PostgreSQL. Dunno about Interbase/Sybase.
Although, I discover that my needs (web backends) are met without using Oracle's bells-and-whistles -- which is why I'll never be a real DBA, just an amateur forever. While I suppose ANY application can be optimized by utilizing Oracle arcana, time-to-market keeps me from dabbling.
But, replication is seriously cool, even for simple web stuff. I like being able to know that my dev DB actually reflects current state (or vice versa, tho that scenario turns my guts to water.. :)
Potato chips are a by-yourself food.
This is not unique to Oracle, but I suppose the argument could be made that Oracle is the leader in this.
This covers mostly the completeness of an Oracle solution.
True.
Not as unique in these days of ODBC. The quality of the middleware tends to be higher with Oracle solutions, however.
Definitely not unique. PostgreSQL, for example, has hooks in Tcl, Perl, Python, C, C++.
I'm not an expert here. I have no experience with I18N solutions in Oracle, but knowing what I do about Oracle as a technology and as a company, I can confidently say that Oracle has best-of-breed solutions in this area.
But, PostgreSQL, with it's world-wide development model, is probably not as far behind as you might think.
No experience here, either. I know that Oracle's embedded Java is highly regarded, but haven't had much of a reason to delve into it (plus, my Java programming is pretty poor)
This has the ring of untruth -- an application that talks to Oracle on Windows most certainly isn't portable to Unix or MacOS. Perhaps the SQL calls and triggers are portable? That doesn't make sense, either, as SQL is supposed to be portable, and triggers are internal to Oracle.
I don't credit Oracle with this anymore than I credit Internet Explorer for rendering <I> tags as italics. Oracle follows E.F. Codd's RDB spec, and as such it allows for multiple users (which, really, is as much a part of it being Unix based as anything else).
This is not unique to Oracle, but is executed well within the environment.
This is a function of an ACID compliant RDBMS, specifically, the "D" of ACID, Durability. Your data should not be vulnerable to some yahoo tripping over the power cord. However, in practice, this is more dependant on the skill and talent of the DBA than the heroics of the RDBMS (i.e. spec-ing mirrored arrays, good backup policy, sane rollback partitions, and proper disk management (rollbacks on separate disk from data, etc.))
Potato chips are a by-yourself food.
I agree that each database product doesn't directly compete. MySQL != Oracle. It all highly depends on what you want to do.
/. where loss of data isn't a concern, but speed is.
MySQL is great for sites like
Postgress is great for those who want a low cost solution for a small RDBMS.
M$ Access is cute for visually seeing the relations, but pretty much a joke for anything serious.
SQL Server / Sybase these two are essentially the same, they were the same product a few versions ago. M$ added some neat features to it, and Sybase seems to be struggling to add useful and needed features. (Like hot backups (where you can take a snapshot of the database devices), load balancing, easier management of replication, major scalability and failover.) Sybase is working on it, but Oracle and DB2 are way ahead these two products by quite a bit. SQL Server is also hampered by only running on intel hardware. Sybase at least runs on quite a few different platforms.
Oracle the kingpin of the RDBMS areana. Pretty much has all of the features, and a huge licensing fee.
DB2 everything that I've read has rated DB2 quite high, it seems to be the database that people use but no one seems to know about.
Pretty much it's all that you're wanting to do with the database, and how much $ you have to determine which database you use. But keep in mind if you write everything you can in ANSI SQL, it should be fairly portable between database vendors. Unfortunatley I've seen cases where writing something with Transact-SQL (Sybase / M$) or PL/SQL (Oracle) is needed. But supposedly both are going to have (or already have) native java support so you can just create a query in java rather than SQL. I don't know much about that other than it's supposed to exist or comming soon, and I'm not a java programmer, so I can't say for certain.
"If you insist on using Windoze you're on your own."
I think there are really three broad categories of issues.
First of all, how important is your data? High end RDBMS systems offer a lot of administrative features that make backing up your data easier and less intrusive. To take an extreme example, Access is notorious for losing data. Everything is stored in a giant blob file, and when that gets corrupt, you can loose everything.
Second of all, how much load do you need to support? Oracle has been running very heavily loaded DB servers for a long time; it's the safe choice. SQL Server is getting better, and MS claims that it will scale indefinitely, but it's probably trickier to run under high loads than Oracle on Solaris.
Finally, what kinds of programming are you going to need to do? Will you have feeds of data moving in and out of the system? Lots of triggers? Integrity checks? How much procedural language programming are you going to have to do, and how closely will you need to integrate the DB into existing systems?
You shouldn't use Access for anything serious. It's a bad system that can't handle loads, and that tends to corrupt and lose data. Think of Access as a 2nd rate front end, and stay away from its back end at all costs.
MySQL is fast, and it's great for simple web applications, but it doesn't support big chunks of important SQL functionality and it doesn't have professional admin tools that you need to protect your data.
MS-SQL is kind of expensive for web applications, unless you're comparing it to Oracle. And I think the development tools are weak when compared to Oracle. But it's easy to admin, it's rock solid, and it has a really respectable features set across the board. MS seems to be intent on closing the features gap with Oracle. Some big companies use SQL Server to store financial data -- it's a reliable platform.
Admining Oracle was hard for me to learn. Setting Oracle up on Linux was painful as well, for me at least. I don't know it well, and I've only played with it, never done anything serious with it, so I'm not a good Oracle reference. But it strikes me as big and complicated and able to do absolutely everything that you'll need. You probably won't outgrow it.
If you're going to run a message board on a web site, then use MySQL and PHP. But if you have data you have to protect, and you need to keep an offsite server that's constantly synced with your main server for safety, you really have to go with something like MS-SQL or Oracle.
You must be talking about the Oracle Applications suite (which is at version 11i now), not the Oracle RDBMS and tools, which are at version 8i.
The article was asking about the RDBMS, not the Apps.
Oracle runs on many platforms. What platform does Oracle run best on?
A: A 35mm slide projector.
(I guess you would have to sit through an Oracle marketing presentation to understand.)
If tits were wings it'd be flying around.
Just like that old phrase says, noone ever got fired for buying Microsoft. The same goes for Oracle.
Please don't rewrite the history.
The correct phrase is "Nobody ever got fired for buying IBM."
FedEx has successful systems using HP/UX and even AIX, but you're right in that if you give the DBAs their druthers, they'll say "Solaris".
And so will all us SAs.
-
For small to medium sized installations of a couple of hundred users Interbase 6 makes a perfectly fine server. It has all the checklist items (transactions, triggers, stored procs, solid SQL support), is a robust, debugged application, and is now completely free. It stores each database with all the users, groups, stored procs etc in a single file, which makes backup and transfer a snap. Newsgroup support is also good, and will probably only get better with time. The management GUI is no Enterprise Manager, but that's a small price to pay for what you get. In certain market segments Interbase could eat MSSQL for lunch if it were marketed half-way decently.
-
Are more than phone drones
- Trust your evaluation of the problem is really exposure of a bug and don't call it user error
- Have ready access to developers and source code and can provide patches/fixes in a quick time period
If any of the above elements are missing -- phone drones who have to run you ragged before escalating to someone with a pulse, fingerpointing about software/hardware "user error", no developer/source access -- you are fucked, and you might as well be posting on Ask Slashdot, since you'll probably get a worthwhile answer there faster than from support. In fact, I know people who will pay consultants when they have problems because the consultancies usually have better contacts and can get beyond level 2 support faster than Joe Support Contract could ever hope to.I don't mean to be knitpicky, but the quote was "nobody ever got fired for buying IBM."
Since you recommended Interbase, have you used both the commercial and open source versions? Now that Interbase is open sourced, is Borland still continuing develop?
I've also heard that there is an Interbase code fork called Firebird. Do you know anything about Firebird?
thanks!
cpeterso
If you have a task that you know you can do with the necessary reliability with MS SQLServer and Access, or mysql, or whatever other low-end database you want... then you don't need oracale. If you have such a task, you also are not a bank doing transaction processing, and you are not doing any truly mission-critical tasks.
Oracle can cost a lot more than the 50 thousand bucks you mention... a lot more. Oracle is extremely scalable, and extermely stable. Oracle offers 24x7x365 support and backs their product 100% (for a fee, of course).
To put it differently, if you were aware of a situation that actually used oracle for good reasons, you probably wouldn't have asked the question in the first place.
Scope: Oracle provides business and enterprise solutions from ground zero to the never ending support and uprgade cycle
Support: You can't really have better support then some of the talent working support. Sure the first person you get in touch with may not have any clue what you are talking about, but the support infrastructure is there and works great! Just create your Itars and TAR's according to the guidelines and recommendations of support and they will fix your issues
Solution: This is the biggy. For example, with Oracle 9i AS you have an incredible applications server that cannot be beat! Built on Apache, Oracle 8.1.7, and Portal 3.0 with Forms and Reports 6i. You can build your own portals running Apache, PHP3, MySQL, and Zope and a conglomerate of modperl and such, but with 9i AS you have a foundation already built that you plugin saving hundreds of millions of dollars in development.
API's & Developer Support: Oracle does provide alot of its technology stack, open API's and such to developers. You buy Oracle Applications 11i and you can integrate that and customize to your hearts content. Most companies wouldn't imagine providing tools like that and supporting it at the same time
Proven Technology: You use oracle software, and your database WILL work, simple as that. Need custom software and you don't want to develop it, oracle will, need an application to solve your business needs, oracle has it.
Open source is great, but commercial software that supports and open environment is just as superb in my opinion. You buy Oracle and you can run it on any architecture available today. You can use the skills and technologies already tried and true in your business to leverage oracle software to increase whatever capacity and concepts your trying to achieve.
Really? This doesn't fit with what I've seen. In my experience, disk read/write performance falls off about the same from first to last sector, on all products in a line. Regardless of the number of platters in the drive.
The fastest drives come out in small sizes first, but that's more of a sales and simplicity of manufacturing I'm sure than anything else. Toss more platters in it and you simply spread the performance curve over a much large data space.
As a reply to the post below this...
It's not as much "See this, it's *real* hardware which you require to run Oracle." as "You don't really need Oracle (or DB2, etc) unless you also need this kind of hardware - if you're playing with x86-level hardware, you don't need the high-end software."
This is where a DBA comes in. They'll know enough about how the data is stored (that stuff nobody is supposed to need to know) that they can tell if page level locking is sufficient. In most cases, I agree, it would be. If you're writing to phone records stored (for instance) in name order, page locking is okay for random access patterns. You'll get a block or two, but who cares in the long run. But if it was St. Patrick's day and you got a a run of calls from names starting with Mc, you'd be in trouble.
This is why keys are often assigned with things like MD5, an even distribution random appearing hash will prevent too many concurrent accesses from blocking.
Then you simply index by everything you want to look up. As long as the table doesn't block, who cares what order it's in.
But a skilled DBA will give you an answer that's related to the actual product you use, not just general comp-sci answers.
Oracle says they saved a billion dollars by using their own software.
Of course, they got it for free.....
The best summary of this is that "The Database IS the Business". It all depends on how much the integrity of the database is worth to you. It also depends on how well you want to scale up the business. And I'm not *just* talking about the database, but the level of support, and the personel who are trained and *qualified* to manage your data.
It's kind of funny to say it this way, but I'd have to wonder what the correlation is between successful businesses, and commercial databases.
At the big business side, an open source database isn't even a consideration. Its at the low-end of the scale that it is. And the single driver for going with an open source database doesn't *seem* to be reliability, cost, or 'openness'. (Do you want your people modifying how the database's internal code works?) Its about dollars. Businesses with dollars will (with rare exceptions) choose a commercial and well-trusted database.
Having a person to answer the fone and write down your question 24x7x365 is only 1% of the job. We've broken Oracle numerous times and it's taken them weeks to get back to us with something to TRY, not even a guaranteed fix.
I think the biggest thing that Oracle really provides is tight row-level locking (Sybase only locks in 4k chunks to my knowledge) and better transactioning or whatever you wanna call it.
Speed too I guess.
You've never used a large RDBMS, have you?
Most RDBMS' do allow you to choose between speed and ACID/atomicity/etc. DB2 supports a boggling array of options, including a MySQL-style "store with no security whatsoever" option. The latest SQL spec even provides standard terminology for some of these features. Most of the big DBs provide more speed vs. reliability reatures than the spec requires.
--
Forward, retransmit, or republish anything I say here. Just don't misquote me.
Oracle is worth it based on what you are doing, I have used MySQL, PostgreSQL and Oracle. I maintain some Oracle DBS at work. MySQL will shine for a news site like slashdot, PostgreSQL would shine for ebay. Oracle on the other hand will shine for something more complex. My DBs handles 10's of millions of transactions a day, very complex transactions, with each DB being half a terabyte, it is no easy feat! MySQL and PostgreSQL will not be able to even handle a quarter of what I give to Oracle. Oracle is way way way advanced, you just don't know much about it that is why you think it is the same as the low end free DBs. Take a trip to a good computer bookstore, and take a look at Oracle books, try reading one on performance tuning and you will quickly realize how complex it is... Anyway, it all depends, on what you want to do and of course how much money you have, and amount of effort you are willing to put to get creative.
------ Curiosity killed the cat. {satisfaction brought it back | it didn't die ignorant | lack of it is killing mankind
I've used Oracle and Sybase extensively. Sybase does not suffer from most of Oracle's speed problems, but neither one supports such features. Nice to see DB2 does, but please don't make sweeping assumptions about my experience.
A more two-fisted approach might be to note that yes, big-ass commercial databases (Oracle) have tons of features that you really, really need sometimes. That being said, the vast majority of sites don't need it. Sometimes, the real database professionals pipe up and see things all from their perspective. Fact is, I can do something the size of /. without really encountering a strong need for real database normalization.
Boss of nothin. Big deal.
Son, go get daddy's hard plastic eyes.
Expanding a vast wasteland since 1996.
Most databases have some sort of an executive that runs scheduled jobs. It would be trivial to implement something which fires off an email message or beeps somebody when a disk is using 90% of it's space. I really don't see why this is such a big deal.
War is necrophilia.
It depends on your app.
Lots of writes postgres is faster then mysql.
Lots of selects with comples queries postgres is faster.
lots of selects with no joins mysql is faster.
War is necrophilia.
Extemely stable as in a 74 pinto is extrememly stable compared to a 72 pinto.
War is necrophilia.
"But where do you draw the line?
"
Here..
If the program is expecting to use some resource it ought to check that the resource is available. This applies for anything. A program ought not to crash because some file it expected is missing and it ought not to crash because there is no more room on the drive. Seems simple to me.
War is necrophilia.
If you are running a DataBase containing a few meg (or even a few hundred) that can afford to come down for backups / maintenance (vacuming anyone?) then you will probably get no advantage from going to a more expensive commercial product.
Oracle (the DBMS I use most) lets you do funky things like online backups (backing up while your database is up & serving data), interesting index options, adding terabytes of disk while the system is up and running. All things that *most* people don't care about BUT are vital for 24/7 shops.
Of course, anything on an MS Windows platform has stability and security problems... but it's not only the platform that contributes to such things. If you're dealing with a database which tracks things that matter, then only a very robust database will do. Use Oracle if you can afford it, Postgres if you can't.
And if you don't have the skills, hire someone who does - it will be cheaper in the long run.
I'm old enough to remember when discussions on Slashdot were well informed.
Is there anything more to this than correct support of Unicode (+ maybe various other national charsets)? I'm having difficulty seeing what a Chinese user would want that an English user wouldn't, besides the ability to use Unicode/Big5/GB/etc..
*All* modern DBs support Unicode, from MS to MySQL to Postgres to Oracle to DB2. But even if one didn't, as long as it supported ISO 8859-1 you could just use UTF-8. Sure, you'd have to multiply your field lengths by 3 to be on the safe side (or 6 if your users want to write Egyptian Hieroglyphics or Klingon), but it would work.
perl -e 'fork||print for split//,"hahahaha"'
In my limited experience, MySQL is faster at doing simple things than any other database out there.
When you're programming a web-based app, where response times are critical, MySQL may actually be a better choice than Oracle, even if you do a hybrid solution, with MySQL used for quick lookups and Oracle doing the back end transactional heavy lifting.
Notice that the Oracle $1,000,000 speed guarantee only applies to DB2, WebLogic and MS SQL Server, NOT Mysql or Postgres.
But you can't. And that's why the big software.
I just came off of a large customer site where, the VP(IS) having received the personal assurances of Bill Gates, the customer spent two years replacing UNIX-hosted RDBMS with NT4-based SQLserver. Result? Transaction query times went from 2 seconds to over two minutes. The UNIX programmers are moved on, the VP is out of the computer industry entirely, and the customer is in deep kimchee.
One of the hardest lessons for apprentice (and journeymen) programmers/analyts to learn is scalability. Just because the single-user prototype worked doesn't mean that the design will scale up to support hundreds of simultaneous users.
One of the marks of the veteran is how fast s/he checks the specs for the marks of scalability. In the case of an RDBMS, if it doesn't have row-level locking with transaction commit/rollback, then it's a toy to be discarded immediately. This isn't because s/he wants to spend the company's money faster, it's because seeing one's software design fold in production like the Tacoma Narrows suspension bridge is a gut-wrenching experience. Those users are riding your design and if they are being hurt, it's your fault.
I won't say that MSQL, MySQL, or Access are inappropriate for all uses; that would be silly. Ad-hoc databases, LDAP, fine. But they cannot hold up in serious, major applications.
Look, you cannot work in the software world without ever dropping a bad design into production. But you can try to minimize the number of times you do so.
And you have a professional obligation to your customer and your users to do so.
I've used PostgreSQL since it was Postgres95 and each major version seems to get much faster, (as in the same query executes in 1/3 to 1/2 the time.) However, I know nothing about running real benchmarks on database systems, so I can't quantify anything. I can only report my impressions, and I think that PostgreSQL is plenty fast enough for use in just about everything you'd use MySQL for.
I'm known to be biased, so make of that what you will.
The problem with using larger drives in arrays, is that the performance will drop noticeably for the same size.
Okay... I'll do the stupid things first, then you shy people follow.
Okay... I'll do the stupid things first, then you shy people follow.
[Zappa]
I use open source (MySQL mainly) databases for about 95/100 projects for which I need database support. That last 5 projects, tho, are by far my largest, for which I use Oracle (and once SQL Server... no more).
The first thing you get with the big boys is customer service support, which is very important, but not unavailable with Open Source stuff.
More importantly, tho, you get third party support. Hardware, backup software, any additional software you may want, are almost certain to support the larger databases, and be efficient about it, too. You're probably relegated to ODBC or NO support for smaller database engines. That's very important for long term planning. Also, if your DBA gets hit by a 747, at the moment it's easier to find someone who knows Oracle than PostgreSQL. (At the moment... hopefully, this will change).
Lastly, the RDBMS supports lots of exotic hardware and OSes efficiently. The difference between Oracle's HP and Sun ports exploit a lot more of the efficiencies of those systems than does a port of MySQL, because Oracle can put a handful of OS specific gurus on the project for each port. Smaller RDBMS may support additional Operating Systems/Hardware, but I'd wager that they rarely exploit them to their full potential. Support, for example, of RAW I/O is available on several platforms from Oracle, while far fewer from other engines. Over time, these sorts of things become important; especially if you ever need to migrate machines, or if performance or large scale is important.
My recommendation would be to look at size, complexity, and lifetime. Larger, more complex (requiring more tuning/speed requirements), and projects with realistic lifetimes of 5 years or more should rely on the big guys. Single DBA projects on budjet machines that really might be replaced in the next few years as people change their minds about things can be made perfectly functional with free (Free?) RDBMS.
If I can take a little different angle than the other posts here, there's more to this than "my product features can beat up your product features".
A large company picking a strategic database may be making a decision that needs to last 10 or even 20 years. Companies that standardized on DB2, dBase, FoxPro, or other older databases many years ago have had to live with that decision. In many cases, they may think they made the right decision and it worked out for them, in others, they may have regretted the decision later, but once you go down a deployment road with a database or any software product, to a certain extent you're locked in.
So, when big companies buy Oracle, DB/2, Microsoft SQL server, Sybase, or any of the big companies, they aren't just buying a technical product, they are buying a certain level of trust in the company that they won't screw them (at least technically, they probably will get screwed financially many times (grin)); that the company will be around in 10 years to migrate the database to a new system; that support will be there when they need it; and that they will have someone's balls to squeeze if the going gets tough.
Open source databases are very, very cool for small web shops and many other applications, but it's not a good fit for the realistic business needs of large corporations. While there is the advantage of having the source code, so you know you can always enhance or change the product yourself, it will be a long time before a typical corporate IT shop is willing to take on that task. They already have their hands full with maintaining less-ambitious in-house applications.
As the comp.risks guys can tell you, there's a lot more to it than just the software.
- "When you want something with all your heart, the entire universe conspires to give it to you" -Paulo Coelho
What OS do you use?
I've used Informix on SGI for 1 year, a 5TB transaction database. And currently Oracle 8i on Solaris, a 100GB datawarehouse full of stored procedures, triggers etc running extremely large (generated) SQL queries, some of the queries alone are 100kb text!!!
In the 3 years intensive working with these databases, I've never seen a single crash, and could not imagine one.
Could the coredumps you see have something to do with the operating system you are using? Such as the ever changing glibc on Linux, which will surely make some applications instable?
>anyway, I wouldn't recommend Access
>for that purpose anymore
Anymore? You mean you _ever_ would have?
james
Absit Invidia
Oracle can do many things that MySQL cannot do. A more reasonable contender against Oracle would be Interbase, which is also open-source, but even Interbase 6.x can be bought!! What you get is a license with some support. The danish banks use this and pay for Interbase 6.x clients.
What Oracle provides is a big framework with education, client development tools etc. The closest I can imagine to that on Interbase would be Borland Delphi, which only now is being released in a Linux version. But the support behind using Delphi and Interbase has no long history and is therefore a bad choice in the eyes of many CIOs.
Lars.
I see lots of comments like "you don't just pay for the database, you pay for the people behind it" etc.
What is that supposed to mean? Does that mean that if you have problems with your database, Oracle will dress up their best monkeys in suits and jet them to your office to charette until your problems are solved? Bzzzt, wrong. That' more money - either per incident, or per service contract.
Do you recieve uptime guarantees? Performance guarantees? "We won't muck up your accounting data" guarantees? Don't think so. Their licensing terms are so chock-a-block full of "we're not responsible, don't sue us if something goes wrong" legalese that I'd be surprised if they ever paid out a penny.
I'm not saying that guarantees of this sort make any sense. I just can't understand what people are talking about when they say Oracle stands behind their products. Only to the extent you pay them they do. Which is fine. But how is this any different than, say, contracting with Great Bridge to support your PostgreSQL installation (to pick an example)?
You get the name, pure and simple. Oh, and you get Larry some jet fuel so he can buzz San Jose in his $38 million Gulfstream V. Hey, it's your money.
I mean no disrespect to the excellent staff at Oracle. But as a consumer, you need to ask yourself what you're really paying for.
--Lawrence Lessig for Congress!
There are some features provided by the so-called "expensive" RDBMSs that current "inexpensive", "free" and "open" source database management systems do not provide.
Replication is a feature that allows databases to keep each other up-to-date. In some modes, replication can occur to allow offsite databases to stay synchronzied with a master. Other modes offer replication, often with only seconds of latency.
Being able to perform transaction log dumps and incremental backups are also a feature the "expensive" databases have and others do not. This is very important to ensure minimal data loss in the event of catastrophic hardware failure.
Clustering, to provide transparent, seamless recovery when hardware fails is also a facility that other database systems to not currently offer. This is important in installations that absolultely require 24/7 uptime.
Some "expensive" SQL databases are becoming particularly good at caching data, query plans and even result sets, so that it need not invest large amounts of resources in oft-used queries.
Some expensive engines are capable of analyzing past queries to determine what indexes and stored procedures should be created to optimize efficiency.
High-end databases have been focusing on providing more than just a tabular view of the data. Many are providing complex text-based searching capabilities, which requires a far-different indexing scheme than what is used in traditional SQL databases.
On other fronts, expensive and inexpensive databases seem to do similarly well, and the inexpensive ones are getting more sophisticated with each release. One in particular that has been making such strives is PostgreSQL.
PostgreSQL still has some quirks, though. One in particular is the need in many situations to "vacuum" the database. Hopefully in a future release such garbage collection and optimization management will be transparent to the user.
My car gets 40 rods to the hogshead, and that's the way I likes it!
The fact is that there are a lot of RDBMS Solutions out there that are a lot less costly than Oracle and offer the real power of an RDBMS. Solutions like SQL Server are far less expensive.
What are the magical features that make RDBMS so powerful? Well things like Stored Procedures to control the input and output of data, Views to control access, powerful Optimizers to make queries more efficiently, Replication to protect critical information from hardware failure, and Transactions to keep atomic operations together. Without these featues we are back to the old days of dBase and storing data in files. It offers no way to plan or tune those operations and no way to protect the data from being used improperly.
While mySQL is fine for simple data storage and retrieval needs it gets out of it's league real fast when you start thinking about a mission critical application running in a high volume environment.
Is there a role for mySQL? Yes! Smaller, simpler sites that don't need to deal with complex transactions or large volumes of access. Is there a role for commercial grade RDBMS? YES! I'll bet you have bought at least one item on the internet through an Oracle or SQL Server DB and you would have been mighty upset if it had failed, or double billed you, or didn't ship you the product.
We have learned a lot from all those years of MainFrame and Client/Server development. Protect you data, keep it safe, use it wisely. You can access it with all kinds of tools from Web Sites to Fat Clients, but never compromise on the integrity of that precious resource!
$30,000 will seem like small potatoes if you orders DB was trashed and there is no fail-over.
9. Performance - simple databases are fast with simple queries, but sophisticated databases are fast with complex queries (and, maybe, slower with simple queries...). M$ uses simple benchmarks to tell their customers that they are fast (TCP-C), but most other vendors have stopped three years ago to post TCP-C benchmarks, so M$ is able to beat three year old databases on three year old hardware... TCP-D is more complex, and M$ couldn't show that their database wasn't 100 (one hundred) times slower on a specific TCP-D part than Oracle - though they tried hard. But benchmarking is a complex matter...
Are you trolling or astroturfing for Oracle?
Go visit the TPC website. Funny how they don't mention the "all important" TPC-D benchmark on their front page. The TPC-D benchmark is listed under the obsolete section under "benchmarks".
Now go visit the TPC-C results page.. The last posted results by Oracle are from January the 18th 2001. Doesn't quite sound like something they don't care about to me. (Although the results are listed as being posted by the hardware manufacturer, anyone who has worked with benchmarks knows it's impossible to post strong results without strong support from the database vendor).
Mmmm.. Donuts
It is *very* convenient to develop for, but something of a pain to maintain. (Depending on wether or not you know your stuff)
I'd say that SQL server works as well as DB2 and Oracle as long as you do not have extreme load.
Of course, again, you must know your stuff. Don't use all fancy wizards and widgets. Script everything and skip the GUI and you'll be fine...
All opinions are my own - until criticized
This topic has been addressed several times already on Slashdot but here it goes again. For 99% of the uses of a database today where data integrity is not of the utmost importance, the data sets being handled are relatively small, and the ability to scale to millions of hits a minute is unecessary, the simple RDBMS's like Access and MySQL can handle the job.
Oracle and DB2 are meant to primarily run on systems that will receive a lot of traffic (both reads and writes) and where data integrity is important (i.e. failure recovery, transactions, integrity checking, etc). Sites like eBay, Yahoo and Amazon are probably better off running Oracle and the like while smaller sites that have traffic that consists primarily of reads and where the integrity of data isn't extremely important can use MySQL or even Access.
Banks and Fortune 500s are also examples of companies that need the additional features of Enterprise database systems. Realistically for most people's needs Access or MySQL are all that is required.
Those are the technical requirements of building a very scalable DB-backed web application. The only two RDBMS systems that truly support these features, to the best of my knowledge, are Oracle and PostgreSQL.
At this point in time, Oracle is more scalable on *VERY* large systems, because it can be tuned more precisely when you have 22 mirrored disks and a 16-processor box. Also, Oracle is usually very stable (although the latest versions which come with a bloated VM and all sorts of additional crud have become significantly less stable than Oracle DBAs were used to).
For most people, though, PostgreSQL is a fantastic technical solution. It provides all the features in an easy-to-manage package with an extremely advanced architecture. Within a few months, it will support replication much more easily than Oracle does, and will continue to lead the way in SQL92 compliance.
Also, PostgreSQL and Oracle both have serious support options. In fact, PostgreSQL is supported by multiple companies, so you actually have a choice. If you're worried about the quality of support you'll receive from these new organizations, then you've clearly never dealt with Oracle's pathetic excuse for support.
There is a catch, though. Your job may be at stake if you use PostgreSQL. No one ever got fired for choosing Oracle. PostgreSQL will probably work better than Oracle at significantly less cost of administration with significantly better support. However, because it isn't nearly as well accepted in the IT community, you are taking a personal risk by choosing PostgreSQL. If Oracle fails, it's Oracle's fault. If PostgreSQL fails, it's going to be your fault for choosing a "non-standard" software product.
From a technical/usability standpoint, PostgreSQL is the way to go. If you consider the politics of your organization, though, and your organization isn't comfortable on the cutting edge, you may still have to go Oracle for a while.
I didn't see any mention in any of the responses yet of the obvious difference, which would be the inability of postgres or mysql to scale to a distributed environment, since neither support multiphase commit. This is the single feature that separates the high-end database solutions from everything else.
maru
Well, don't get me wrong, I love Postgres, but I doubt it could handle the transaction load of eBay. Further I doubt it could run on hardware robust enough to not instantly croak under the load (eBay's peripheral machines like webservers are all NT/2K (dumb IMHO), but the central database last I heard was pure oracle on solaris (and presumably one to several big Enterprise machines); I guess you could compile postgres on solaris, but would it be able to take advantage of the features of something like an E10K?).
--
News for geeks in Austin: www.geekaustin.org
News for Geeks in Austin, TX
For the love of all that is decent and holy, you get a code shop that can afford a decent Quality Assurance process. Linux the kernel is fine, but too many things ship as part of a system that are there because they compile with nothing worse than warnings.
Vintage computer games and RPG books available. Email me if you're interested.
GemStone/S is in version 5.1.4. I would hardly call it a "research idea." It is a high-stability, high-performance, and costly DB. Get over it, man.
Working toward a usable PDA environment in the spirit of Newton OS: Dynapad
Hmmm..., not *all* websites are small. Think about the Avis, American Airlines, or Best Western websites. Don't you think those sites are mission critical and hold vast amounts of data?
I agree with you, in that I don't feel ready to entrust one such system to mySQL.
But, although I'm still trying to muster up the courage (is that the correct way to use the verb "muster"?) to use mySQL for a large database, I have already taken a large step in getting rid of the large-database guys. I recently had to upgrade an Ingres database system. Rather than pay $30000 for an extra licence for the back-up computer, I got a fully redundant machine, with dual power supplies, 3+3 disks RAID, etc. I hope someday I will be macho enough to stick a finger to those guys at CA and use a free database...
Why is it that the proponents of "one nation under God" are so eager to get rid of "liberty and justice for all"?
Although I'm just a mere home user, I've contacted various OSS developers on problems with the software they wrote and got the exact reply I needed to fix it. I've been in IT for 12 years and have dealt with the likes of IBM, Compuware and CA, and despite having problems with their software the answer varied from disinterest to slightly helpful. You need to be someone like Citibank before corps give a damn, whereas the OSS guys prefer their software to be working for as many people as possible.
Although DB2 on a mainframe absolutely creams Oracle on Unix, so is the only real option for large-volume batch processing. Rexx/DB2 is just as good if not better than PL/SQL and
DB2 has had stored procedures since v4, but I'm not sure what the other 2 are though, so maybe you could clarify.
I'm developing programs for financial institutions. Here are the advantages of commercial solutions over the free ones. That said, YMMV. 1. Support - you get support around the clock from the commercial vendors. This goes so far that, for Informix for example, the actual source code is handed every 8 hours to the next team that starts working (so the source code wanders around the world in 24 hours). 2. Transactions - yes, they will add transactions to MySQL in the future, but for now I go for safety. The database vendors have years of experience, and considering how complicate it is to do reliable transactions, I'll go with the experienced ones. 3. Scaleability - you get the commercial databases for nearly every platform that you can lay your hands on. If your database is too slow, you can always change your hardware (this is the major reason why I don't think that SQLServer from M$ is good solution for my needs). 4. Migration - with commercial vendors, you can always upgrade to the next version without pain (it killed Sybase nearly when they didn't do it right once). You can easily migrate to other platforms (but not necessarily to another RDBMS). 5. Standards - only the big databases confirm very rigid to standards (though they all try to lock you in by offering enough "nice goodies", and if you're dumb enough to take the offering, you're lost). For all major database vendors you get ODBC, JDBC, native drivers etc. 6. Interoperability - if you work with tools like ER/Win, development tools like Prolifics Panther or Power Builder, they are supported. Don't count too much on ODBC, its slow, and it does not support everything you need. 7. Multi-tier - if you need to support many , many users, you will have to use n-tier environments, and the transaction monitors (Tuxedo, JetNet etc.) will support only the major database vendors. And one day you'll wake up and need just that. And with the internet, this day will come earlier than you think. 8. Database replication, load balancing, fail-over and such things are, to my knowledge, only supported by big database vendors. It took them more than ten years to get this right (and M$ is still not able to do it right, though they claim the opposite). 9. Performance - simple databases are fast with simple queries, but sophisticated databases are fast with complex queries (and, maybe, slower with simple queries...). M$ uses simple benchmarks to tell their customers that they are fast (TCP-C), but most other vendors have stopped three years ago to post TCP-C benchmarks, so M$ is able to beat three year old databases on three year old hardware... TCP-D is more complex, and M$ couldn't show that their database wasn't 100 (one hundred) times slower on a specific TCP-D part than Oracle - though they tried hard. But benchmarking is a complex matter... 10. Added values - like Object-relational databases - are a speciality of Informix and Oracle. If you need this, there is not much choice. E. g. storing, retrieving and searching of word-documents (with indexing), is it possible with MySQL? No. 11. OLTP - in this area you have no choice but use a big vendor. If your database has to run 24x7, can you do a reliable backup on tape or anything? Can you restore parts of your database while the database is online? Reliable backup means: a transaction-safe backup which will make a snapshot of the database where every transaction is in its correct state, and a continously running backup. 12. If you ever need data mining etc.: look what databases are supported. 13. If you have a few TB to store, you can't use MySQL. It might be possible, but I doubt that this will be fast. Most simple databases don't scale with sheer size. I can only stress that reliability has its price. And there is more to reliability than just transactions.
You found a sword: +4 damage, +5 moderator points
I think the decision betwene mySQL v. Oracle or MS-Sql is the same as other opensource v. non-opensource decisions. Here are the eternal truths I see on the topic...
...
1) There's always a faction that will go with Opensource just because it's opensource. Usually the underlying reason is philosphical: "I don't want to sell out" or the like
2) There's always a faction that goes with commercial software because its commercial software. Usually the underlying reason is "it's not real software unless I pay money for it"
3) The majority of people make educated decisions between the two options on an individual case-by-case basis. Usually, these decision are based on answers to the following questions:
- do I need extra features?
- do I need support?
- what sort of software does my shop currently use?
- what sort of training is involved?
- what sort of specialzied hardware or software is needed?
- how much installation time is needed?
Then, it's been my experience that people's biases are factored in after answering these questions. For example, I would like commercial support, but I really don't like Microsoft. So I chose an open-source solution.
I think, as a general rule, commercial software is going to be more functional and backed by better support (with notable exceptions). But, if you don't need the support or you don't need the extra features, then why bother paying all of that money? I think a case-by-case decision is the best way to go.
GnuPS ia a new SourceForge project to develop a peer-to-peer parcel delivery service.
Parcels are carried in hub-and-spoke fashion by opensource developers commuting to and from work in major metropolitan cities. The service is open to anyone without any restrictions, except that all parcels delivered by GnuPS must be available for inspection by anyone.
Although delivery options, service time and reliability don't yet match that of commercial parcel delivery services, GnuPS is improving every day with suggstions from GnuPS users and new volunteers in the GnuPS delivery network.
GnuPS runs on Linux 2.4.2 + MySQL 3.23.33 + PHP 4.0.
GnuPS CentralOffice requires a minimum of a 486SX-50 PC's with 32 MB of memory, 400MB disk space, a CueCAT barcode wand, and an interface to GnuPS Drivers (via 10BaseT or a docking cradle).
GnuPS Driver requires a minimum of a Linux-capable handheld PC or laptop, 30MB of storage (disk or flash), and a CueCAT Decoder.
you are building a website. You might take a look at http://www.onlamp.com. Has some good info. Granted it leans towards MySQL but it does explain why so many websites use MySQL. Having said that if you need *really* big databases Oracle does rock hard.
Cypherpunks: Civil Liberty Through Complex Mathematics. Those who live by the sword die by the arrow.
In life there are always extenuating circumstances. Unfortunately, bosses who want blood from a stone are quite common.
I bet when you called Oracle support, you didn't have a chip on your shoulder, some people do. Hell, I corupted a database once with a bad script. At first I thought it was a problem with the database, because I looked over the script and couldn't see where my mistake was therefore there was no mistake. When I called support, they were a little skeptical that it was a bug and instead wanted to see my script. They took one look at it and politely tore it apart. They told me how to fix the database, how to rewrite the script, and to call them if I had any problems. It was a humbling experience, because I was to blame for my mistake. Some people can never admit when they're wrong and support people have to deal with them daily.
I only hope that when you did screw up, your bosses didn't hang you out to dry. I've seen that happen many times before. A few words of advice, suck up as much knowledge and experience you can and find someone who is willing to pay you to be a real DBA.
BTW, Oracle support is the best support organization I've ever encountered, you really can count on them.
"The words of the prophets are written on the Slashdot walls."
The problem with your thinking however, is that no large scale application or even small scale for that matter can trap for all the things that can go wrong in an operating system.
The last thing I want is for Oracle or any other DB vendor to spend so much as fifteen minutes developing error handling to account for things any brain dead sysadmin should already be monitoring with at the very least a cron job. It is simply not their responsibility to check for every event under the sun.
"The words of the prophets are written on the Slashdot walls."
Technically no it isn't a relational database. According to Codd's definition of a relational database the software must meet a certain set of criteria.
- All information must be represented explicitly in one and only one way: as values in tables.
- Each and every datum in the database must be accessible by specifying a table name, a column name, and a primary key.
- Null Values must be treated systematically.
- At least one character-based language must be provided which can be used to modify the structure and contents of the database. This language must be able to do all of the following:
data definition
view definition
data manipulation
integrity constraints
- At least two integrity constraints need be supported:
No primary key can have a null value, and no primary key can be duplicated.
For every non-null foreign key there must exist a matching primary key.
Although non-relational tools are allowed, it must be possible to manage the database using only relational tools, and the non-relational tools must not be allowed to bypass the integrity constraints imposed by the relational tools.
The last I heard, MySQL did not support foreign keys natively and while referential integrity could be maintained through code, the database itself cannot insure RI.
However, I believe whole-heartedly that this omission is what makes MySQL so fast and ideally suited for the web.
"The words of the prophets are written on the Slashdot walls."
But where do you draw the line? While I'll admit that there are certain circumstances that should be checked for there does come a point where the application vendor must say, "This is the users/admins responsibility. Without that dividing line between application and user responsibility, then all applications will eventually become operating systems.
In the example the poster stated, those admins are clearly derelict in their duties. Every admin worth their weight in dung has some kind of monitoring in place for that kind of thing. Be it Tivoli, Shell scripts, Oracle Enterprise Manager, or even checking the space manually every so often would have clued them in ahead of time. And even if they somehow missed it, a competent Oracle DBA who stumbled across a hung database would immediately check for a log switch, missing file, or lack of free space. For DBA's this knowledge should be as second nature as breathing.
If John Doe goes out and buys a Porsche with a five speed manual transmission without even knowing how to drive a stick, is it the car's responsibility to tell him that he needs to push in the clutch when he comes to a stop? No it isn't, because there is a modicon of user responsibility that they know what they are getting themselves into.
"The words of the prophets are written on the Slashdot walls."
First of all, what is the DB supposed to do. Set off a klaxon horn? Call the hardware vendor and request another drive? or just start overwriting files at will?
I don't think you understand how Oracle works. One of the reasons people put such a premium on Oracle is because of its reliability and durability. Oracle (in an optimal configuration) will back up every transaction log it writes, if this back up does not complete and the database continues on its merry way then the DB will most likely not be recoverable if it should crash. As a result Oracle has a brake mechanism built in that if the database for whatever reason cannot backup the transaction log, switch log files, or write to a datafile it is going to stop in its tracks immediately, because at that specific moment the database is recoverable. If it continues it may not be. This is not only acceptable behavior for a DB it is desirable. Most people would rather have the DB come to a grinding halt and do absolutely nothing until the "skilled DBA" has time to correct the problem. And if that "skilled DBA" was doing their job, they got a page as soon as it happened, if not before. I personally would not want my DBMS do anything more than that, I'll happily leave the rest up to the monitoring scripts, Enterprise Manager Event handler, and the cron job running df every other hour.
"The words of the prophets are written on the Slashdot walls."
Well, I don't know about mySQL, but GreatBridge provides 24x7x365 support for PostgreSQL, which is also open source and said to be more scalable than mySQL.
1. Online backup. The database system that I manage needs to be up 24x7 and needs a 100% reliable parallel backup system Oracle, Informix, etc offer that. El-cheapo DBMS's do not.
2. Referential Integrity and Triggers.
3. Documentation. Commerical RDBMSs include accurate, up-to-date multi-volume documentation sets.
4. Storage management. If I stick data into my database. I know that it is there and accessible. How do I know this? My RDBMS spoke to the storage hardware directly. A glorified flat-file DB like MySQL does not give me the confidence.
5. Scalability/Flexibility. I don't know about Oracle, but Informix can scale from 1 to 64 CPU's over 1 to 128 servers.
Right now we run Sun Enterprise servers. If we decide to migrate to DYNIX, HP-UX, Windows 2k, Linux, etc we can without worry. You cannot do that with MySQL access or Postgre. (since you have to worry about incompatibilities between filesystems)
6. When we find a bug or have a serious problem, we talk to skilled engineers on the phone. No trolling newsgroups or modifying code by myself. I'm a DBA, I don't have the time or the desire to hack database code. We pay good money for support, but then again, the database makes us alot of money too.
If you suggested running Oracle or some other expensive RDBMS for a little website, I'd say you are crazy to pay that much. Same thing goes for a small inventory database or something similar.
At the same time, if you run a large multiuser financial, sales, or data warehouse application on a 'free' database you are far too reckless and incompetent to be a IT professional.
Conformity is the jailer of freedom and enemy of growth. -JFK
mySQL doesn't do lots of things, like 2-stage commits, rollbacks, remote replication. Access is a horrible excuse for a "database" and shouldn't be considered for use by more than one user at a time, ever. Postgres and other open-source RDBMSs are closer to having the "full feature set" that the commercial vendors advertise, but are still a couple of generations behind the technology curve and can have some serious support issues if you don't want to grow your own internal database hackers (which most companies don't want to do).
Then you get into high-availablilty, massive data storage requirements, huge transaction-per-second rates, complex apps where you want to enforce rules in the database... these are the areas where the commercial databases compete with one another. Oracle, Informix, Sybase, and SQL Server are the big names we consider most when we start thinking about a project, although IBM keeps trying to make DB2 a competitor. For personal and small group use, heck, PHP and mySQL works great. For Amazon.com, it's a different story, and for the back-office types that crunch lots of data lots of times (insurance companies, etc.) mySQL doesn't even show up on the radar.
I love vegetarians - some of my favorite foods are vegetarians.
That being said, part of the support issue in that end of the market is design. You get an awful lot of adhoc awful designs. An example would be using a single name field. Not separate fields for first name, last name, title, etc. - but a single field for everything, entered any old which way. Great for mailing labels, useless for anything else. Stupid design, as as soon as you have thousands of records, and you want to find an individual person, etc. And then you have somone who calls up, and wants to sort this out automatically (HA HA HA!)
One call I recall was one person who wanted us to help him design his invoicing system. I asked him if he understood accounting. He said "No ..." I asked him if this was his first computer ... (you can figure out the rest.) Point being is that some of these things are complex, and they are not intuitive, and there is a market for some sort of solution.
Another part of the problem at that end of the market is the new owner of the business. A guy sets up his files with the old dos program, learns it pretty well, gets to be a wiz, grows the business well, and then decides to sell the business and go play golf or something. The new owners do not know dos, do not know database basics, do not know that design or that system, and might be simple users.
And now the madness begins, when they want to move it to a windows system, and do not know the windows version either.
The mom and pop shop are usually farming their work out as far as a database system goes, or they are band-aiding it with what they have been sold on as a solution, such as that intuition impaired product of the Very Important Software Company. Let's face it, that was designed for people who are programmers, and even then it can be a rather big hassle.
The best low end rdbms I have seen is the one put out by Alpha Software, at least for the Windows platform. Granted you still have to know something, but you can get something productive on it faster than anything else out there I know. They used to be a bit bigger, but then there was the power of Microsoft marketing.
There is a market for a good solution, but sadly, the dominant player is putting out something ordinary folks cannot use. You have to be a guru.
"It is a greater offense to steal men's labor, than their clothes"
Mission Critical does not always equate to vast though. Large and critical in these cases, but still not vast (only in comparison to something small... think multi-terrabyte databases)
---
Video meliora proboque deteriora sequor - Ovidius
Yes it is a demo system, but it runs real applications and a full support load (simulated customer service requests, live data warehousing, etc.). It was originally setup as a Telco ZLE, and operates continuously with a simulated load (90 days of up to 5 times the current number of phone calls made world wide). To quote from the website:
The benchmark implementation retains 100 billion Call-Data-Records (CDRs). At peak operation, it accepts up to 50,000 random transactional inserts per second through the CORBA-based application and handles more than 100,000 customer service agents performing more than 3,000 transactions per second in the TUXEDO environment. It also processes 100 customer profile updates per second, physical defragmentation of CDR tables, trickle feed to the data mart, and batch extracts to the mining cluster and the massive ad hoc parallel query.
I would be interested in hearing about other large databases (business, not scientific which tends to have a very different set of db requirements). I know it was quite tricky designing a database that can handle data of this size, and still deliver acceptable performance while mixing inserts, searches, and parallel queries.
as a programmer who has worked with mssql, I have to second that- MSSQL does not scale that well, and even worse, it has all sorts of nasty problems with memory leaks and very awkward allocation schemes. MSSQL will blindly cache 10GB result sets in 1GB of memory while waiting for the socket to catch up and send it all to the client- it will then refuse to give that memory back to the system, or even for other queries. For trivial implementations, MSSQL is a pleasure to work with, but it is a total bagbiter when you need it to do heavy duty stuff. I havent had any problems doing transactional stuff on mssql.
I think my company might have switched entirely to linux/oracle if it werent for oracle. Why do we use MSSQL instead of Oracle?
-cost: oracle costs a ton more. If anything, oracle is a worse leech that microsoft ever was.
-administration: oracle admins cost a lot more than mssql admins. The idiot console (enterprise manager) is easy for anyone to learn really quickly.
-administration: its a lot easier to hose an Oracle installation if you are a moron and dont think carefully before you do something. Did i mention that proper Oracle admins cost more? Hehehe.
I know that this is going to be flamed and marked down as a pro-microsoft troll, but it isnt. I hate microsoft as much as anyone else here. However, is Oracle really that different from microsoft? Hardly the lesser of too evils. I wish that postGreSQL or mySQL would get its shit together so that I could have a COMPLETE open enterprise setup. This would save on licensing and peace of mind.
I work in Support for an RDBMS vendor. And before I did that I was a DBA for several years.
24x7 support is never to provide bug fixes. It's to give you emergency help to get your system back up when it's down. The timeframes for most bug fixes are days if you're lucky. 24x7 means that there's someone at the end of the phone who has the experience and confidence to talk you through getting your system back.
9 times out of 10 it's not that difficult either because there are a limited number of options open to you. Usually you'll be restoring a back-up or undoing a configuration change that had unexpectedly disastrous results.
As a DBA the nice thing about that kind of support call was there was someone to share the heat. When the production systems crashed - as all systems do on occasion - there would probably be 3 or 4 managers standing behind my desk asking me when it was going to be back. It was great to be able to say "I'm on the phone to Support now"
As for the "user error" response - it does happen I'm afraid but let me say this. I get so many calls from people who are supposedly IT professionals who have no clue at all that it's real hard to stay un-cynical - but I try. But I also make sure that I get them to check the simple things first. Sometimes people hear that as patronizing or me blaming them but it's not - it's a troubleshooting procedure. I know for a fact that for every guy who's offended by me asking if any of the filesystems are 100% - there's two who haven't checked it. (And 5 others who just say no and we move on.) Believe me we'll quickly get to the non-trivial problems if that's what it is.
I'm not saying all support is wonderful but I know ours is pretty good. I'm not saying Open Source support is bad either. I do think just having the source is not enough though. You need someone with the time and skill to actually do something with that source - whether that's in-house or a 3rd party. Also most problems I encounter don't require bug-fixes - just knowledge of the DBMS configuration and commands.
I think looking at the quality of available support is an important factor in choosing an RDBMS. Availability is usually not a factor on a commercial RDBMS - but quality may be.
Having said that I think features is probably a bigger factor. From what I've seen so far the available Open Source offerings are still playing catch-up on features.
Not only DID we pay this ammount for the pentiumII install, but I used their web site to generate the other figure of the ppro180. Keep in mind this is for a web site and not a developers license. Thats how they getcha.
Sneakemail is to spam filters what an ounce of prevention is to a pound of cure.
That a very reasonable price and very un-oracle like. I've priced a 2x ppro180 and my company bought oracle for a client for a dual 400 pentium II, both for a public web server, and both prices were about $200,000-250,000 dollars. Those prices make any alternative look good.
Sneakemail is to spam filters what an ounce of prevention is to a pound of cure.
I think there really was once a real "peer-to-peer" public mail service started by someone back in the 70s or 80s. I seem to remember that he was killed and someone speculated that government agents or post office staff had killed him. I've lost the bookmark and can't find it in a search. Does anyone know anything about this?
--
Libertarianism is rich wolves and poor sheep playing gambler's ruin for dinner.
One of the great things about Open Source/Free Stuff -- and I think one of its broadest appeals -- is that it's easy to learn how to use it and what it's about and what it can do w/o shelling out much other than some spare time.
I'd love to learn about Oracle and how it works. However, I think most certification classes are lame, and don't want to pay for them. I can't to it on the job because I can't seem to get a job as an Oracle admin until I know my stuff (fair enough). And I can't spend $10,000-$30,000 on an installation.
Is there any (legal) way to get a hold of the software and begin to play with it until I understand it?
--
Libertarianism is rich wolves and poor sheep playing gambler's ruin for dinner.
MySQL also supports an annoyingly incomplete subset of the SQL spec and of our (and my) required features. I need transactions. Period. (Please, none of the "transactions [as well as unsupported feature X] are unnecessary" arguments that proliferate everytime the MySQL debate pops up on /.). I needed row locking when MySQL didn't support it. Period.
Today we run on Oracle on Sun servers. Oracle doesn't have alpha transaction support. Oracle on Sun servers doesn't have scalability problems. Oracle offers proven, out of the box failover support. Oracle brings a DBA with letters after his name who knows all the dirty little corners of Mr. Ellison's weathered and polished computer program.
But all the developers still run their development systems on MySQL, 'cuz the Oracle client sucks ass :)
I would recommend using what works for you and what conforms to your business. If the database is mission critical, then having the 24/7 Oracle support may be useful. Oracle also likes to tout how they are faster than the compeition, thus the whole 1 million dollar bet they have. However do not overlook MySQL, as it could be exactly what you need, plus the added advantage of it's cost. Negative being no formal support infrastructure(That I know of) along the lines that Oracle has. Prehaps try setting up a test server with MySQL and do some load tests on it that you believe you be fairly close to what your production server would be handling. Actually, the old phrase goes "No one ever got fired for buying IBM". IBM was around LONG before Microsoft.
No need for evaluations and proof of concept trials. Just look at this site (/.) with all its hiccups and outages and you realise why you want to run Oracle. Oracle software powers the internet.
Your pizza just the way you ought to have it.
The issue is obviously multi-faceted. The poster doesn't mention what size dataset the actual DB will be used for. Are we talking about a departmental server, a back-end for a small web application, or a DB to track airline reservations? Once that is known, the issues become less cloudy. One wouldn't attempt to run an airline reservation system on MySQL but they might run in for a myriad of other applications.
The other issue is hardware. Most of the "Big Iron" machines don't run Linux or *BSD because the companies have developed their own OSes. Clustering may help here though.
Back to the main point, cost. The only issues are the return expected and support. If you are running on extremely slim margins in your business, and who isn't, and if your developers are skilled with MySQL, then by all means you should use it. On the other hand, if you expect the DBMS to pull in 6-7 figures a month in incremental revenue, then maybe you should use Oracle and get Tier 3 support.
Value based analysis will tell you which one to use. Using one over the other blindly is just foolish.
__________________ Hey Moderators!! Fuck Off! Thanks.
I can't really speak from experience, only from what I've heard from friends and co-workers, but there can be a benefit of publisher/developer support vs. 3rd party support.
If you have a problem that the support people can't figure out, there's always the possibility that they can contact one of the developers directly, thus getting results. (Either a patch or a workaround somehow) Now, I know that this same thing is possible with Open Source, but if the company doesn't have their own developers, then you have to wait for one of the developers to fix the problem. (If you [i.e. the company] can't fix it yourself, then you probably wouldn't be calling tech support])
It's nice to have things free, but sometimes you need things *now*. Of course, if the developing company doesn't do that sort of thing for you, then they're certianly not worth the extra money. (Unless they are much faster/more stable, which I doubt.)
"The circle of spice, lying on the ground... who put it there?"
Given you even mentioned the multiple thousands of dollars it actually takes to get into a real Oracle system I shall assume you or your organization can afford it. I do have to say this. My wife is one of those people who carries the cell phone 24/7/365 as an Oracle employee. And she is good, real good. And I can say that she does get called at 2:30 am and she works on the problem until it is fixed, and not one minute less. Nothing else is satisfactory with her management. So if you do buy that Oracle system and you're up hacking away at 3am and just cannot get something to work and you've got to have the quarterly results out by tomorrow afternoon, you'll probably be waking my wife up. And from what I've seen so far, your numbers will probably go out on time. Just don't ask me for help because I'm probably just going to roll over and moan. Plus I personally know nothing about Oracle.
Caution: Contents under pressure
I've read through a number of replies and people seem to be concentrating on the support contract and scalability. Quite right too - these things are important. However, it should also be noted that mySQL in particular, which most people are referring to as an open-source alternative, is feature incomplete. It does not support foreign keys, nor does it support stored procedures. It does not support correlated sub-queries. It does not support triggers. On the web page, The mySQL team argue that lack of foreign key support is an advantage. I simply cannot agree with this - working with declarative referential integrity can be more awkward, but as the name imples it is the best way of ensuring logical consistency throughout a database. mySQL has its place, and I -am- a user of it. I wouldn't swap out Oracle or Sybase however and, to be fair, the mySQL team don't really seem to expect me to. Cheers, Ian
This is a laughable question to anyone who knows these products. In certain situations, Access or MySQL will outperform Oracle, even when Oracle is running on a million-dollar E10000. There's no way that Oracle can match what is essentially a flat-file read in one of these little databases.
BUT THAT'S BECAUSE ORACLE IS DOING A LOT MORE WORK!!! WORK THAT SERIOUS ENTERPRISE USERS NEED!
I have used Access/MySQL extensively (and successfully) for many years, and I currently use Oracle. In these Oracle database, my company maintains medical records in several 250GB+ databases (almost 1000 tables each) that at any given time have hundreds of users.
We would gladly stop transferring zillions of dollars a year to Oracle Corp. (and enriching Larry Ellison, that horrendous asshole), if you can give us a matching small-database solution. Sadly, we would be lost without Oracle. You ask what the advantages of a high-end RDBMS are - well here it comes: Does Access or MySQL...
1) Have many different index types to facilitate different types of joins and database uses? (Star queries, bitmap indexes, etc...)
2) Have a serious security model that is extremely fine-grained and capable (and correspondingly annoying to work with, of course)? We work with medical information, and unlike almost every other company out there today, we guard it extremely tightly.
3) Have fine-grained locking? I know the MySQL guys are tired of being pounded on because their locking isn't row level (I do know that they are improving this). But it is important when you have hundreds of users. And just as important is analysis capability: when users get locked, Oracle can tell us (with a little bit of sweat) what exactly the problem was, so that we can maybe fix it.
4) Have a great cost-based optimizer? Is the optimizer flexible enough to respond to hints for those rare situations when the programmer knows better than the database?
5) Have full-featured replication?
6) Have support for numerous production support tools that can report on time-consuming SQL statements and also do things like rebuild huge indexes without taking the database down?
7) Are completely multithreaded so it can take advantage of multiple processors? (Yes - most databases are IO constrained, but when you have powerful disk subsystems like we do, IO can increase to the point where CPU can be a bottleneck.)
8) Can horizontally or vertically partition huge tables so that they are spread across many different disks? This vastly increases IO.
9) Have a flexible and intelligent constraint system that has enterprise-level features such as the ability to enforce constraints on new rows but let pre-existing rows violate the constraints. Sounds stupid - but when a table has tens of millions of rows, it might take weeks to write the program to fix bad data - but in the meantime at least you know that no new bad rows are being created.
I don't mean to slam the smaller databases. In 95% of systems, they are probably OK. But for us in the remaining 5%, our databases are everything to us, and we'll pay almost any price for integrity and performance.
Atomicity. Either everything in a transaction gets committed, or everything doesn't. You make sure that no matter what happens, with anything, if you shipped one widget, that inventory is updated accordingly, etc.
Consistency. The database goes instantly from one valid state to another. It never allows integrity constraints to be disobeyed. Never.
Isoluation. The database gets updated in a manner such that until a transaction is complete, no other query shall see any portion of the results of that transaction.
Durability. Once committed, the results of a transaction are permanent. If the database returns success, then 2 milliseconds later a hole in time and space sends one of the db hard drives to 5th century china, after inserting a new drive, the state will be exactly as it should be, with no information lost.
MySQL, for example, fails every single portion of the ACID test. Yet people will tell you to run your business on it because "it's really fast!" Of course it's fast, it's the db equivalent of a honda burning nitrous, with no way to turn the nitrous off. It'll be fast as hell, but you know it's only a matter of time until it blows up.
A lot of free databases lack basic capabilities such as sub-selects. This is where the value you're searching for is actually the result of another search. Yes, it's fully possible to write code which will do this in the client, but that's ludicrously inefficient if you're working with moderately sized datasets or larger. Instead of letting the db take care of that natively, you not only do the work yourself, in a non-standard hack, but you pass large amounts of unused data across a network stack, which uses resources.
I'm not saying all free databases are evil, they aren't. PostgreSQL is a very reasonable database which passes 'ACI' out of 'ACID'. That might be good enough for a lot of applications. Hell, flat files are good enough for a lot of applications.
The key here is to evaluate what you're doing. Before you look at the cost of the database, look at the cost of downtime, the cost of lost data, the costs associated with misreported data (caused by, for example, the database updating non-atomically), etc.
If the answer to the above cost reporting questions are all 'zero', including salaried time, then use MySQL. If not, consider using a real RDBMS.
Contrary to slashthink, businesses aren't completely stupid, they don't spend large amounts of money unless there's the benefit is expected to exceed the cost. There's a reason why most every business on the planet uses a real RDBMS.
--
"Don't trolls get tired?"
You see, when properly setup, most databases keep the information written across numerous drive arrays. If you follow, as an example, Oracle's recommendations on how to setup your hardware, you'll never have data which cannot be reconstructed after the loss of a single drive.
Yes, almost all software will let you do stupid things like use files instead of raw partitions, or to use hardware which allows unrecoverable failure. The question is whether such a system is "properly configured" according to vendor documentation. In the case of commercial DBMS, such as DB2 or Oracle, they aren't.
Personally, I'd be a little upset if somebody claimed a database passed the 'ACID' test, but had inserted a notoriously unreliable single point of failure (hard drive) into the system. Wouldn't you?
--
"Don't trolls get tired?"
That's the extreme case; in rather less extreme cases, a benefit of Paying The Bucks for Oracle is that you can do online backups while the system is running, and have the expectation that any DBA that is worth his (or her) salt can bring it back up if there is a problem.
Those are the critical things that cause managers to be willing to pay the Big Bucks for licenses.
The fact that there's a sizable community of DBAs from which to hire is a nice bonus; the process feeds on itself in that it is profitable for everyone (well, almost everyone...) to get involved in certification activities. That means that there's a half-reasonable lowest common denominator set of skills out there.
Another "merit" is that there is a whole pile of third party software out there, whether in tools to help you manage databases, tools to help you build DB schemas, applications running atop Oracle ( e.g. SAP R/3 ), and such, again, with financial benefits to many of those involved. It is SAP's interests to promote Oracle sales, and vice-versa... (Albeit that being an example of a situation where there are some conflicts of interest!)
If you're not part of the solution, you're part of the precipitate.
High-end RDBMs (Oracle, MS-SQL, DB2) are beasts of a whole different nature from your regular, run-of-the-mill MySQL or Posgress installation.
First they have a price tag attached. A very high-end price tag. The figure you give for Oracle is the barely minimum a company would expect to expend at the lowest entry-level.
Second, they all must be baby-sited. You do not install Oracle from a CD, make it run and leave it alone. No, you musrt first pain-stakingly install it, sweat and curse for a couple of hours to make it run and them administer it forever. Leave it in the hands of the users for a couple of weeks and the meaning of the word entropy will soon be known at the upper organization levels.
Allright, these above are a bit dramatic but true. In a sense, RDBMS makers and specialists are fighting hard to defend the last outpost of what Allan Kay once called "the high priesthoood of a low cult" (talking about the computer people of the 60's).
Now, high-end software is for high-end jobs. First, they have a bunch of weird performance marks to show at Transaction Processing Performace Council results for TPC-C, for instance. Carrefully note the kind of hardware they are talking about. They are measuring RDBMS performance in machines costing millions of dollars!!
And then these beasts are capable of performance peaks unheard of in the free packages. Not that I do not like MySQL. I like it a lot and have used it in many enterprise systems I developed. But those were small enterprise systems. They did not had to sustain high transaction periods, nor high connection counts, nor gigabyte sized database tables. And certanly not all three together.
I am currently working in the development of a system for a large laboratory. They process thousands of clinical exams per day, mostly in the morning. The organic material for the exams must usually be processed within two to four hours after it is collected from a patient. We will pay whatever it takes to have peace of mind. And I tell you, those will be some well expend millions of dollars.
One thing that is different between commercial and free solutions is support.
I know that Oracle can offer you 24x7x365 support options if you have enough cash.
Can mySQL do that? My experience with support on open-source solutions has been good, but not good enough to bet my business on getting a problem fixed in a few minutes instead of a few hours or days...
Still, sites that use these databases are still prone to problems, just look at EBay!
-Hunter
RateVegas.com - Vegas Reviews
With a commercial database, you pay for proven reliability, proven scalability, support, thorough documentation, and features.
MySql is not a database in the same sense that Oracle is. The four basic criteria that need to be met by any database are summarized by the ACID acronym (Atomicity, Consistency, Isolation, Durability). MySql satisfies none of them and lacks a huge number of very basic features. It is not a database, but a simple SQL parser on top of a DBM file.
The ACID test assures that the data will always remain in a consistent state, despite failures. Most enterprises (banks, businesses, etc) would pay any amount for this feature.
Postgres is a real database. However, it is still _very_ slow and scales poorly. The documentation for Postgres has glaring omissions, and the database lacks some features.
It is often worth paying for a commercial database. There is not a substantial difference between Oracle the "other" commerical databases (DB2, Informix, Sybase, etc). All of those databases are enterprise-class and offer features and performance similar to Oracle. Oracle charges 3-4x as much money because they are the market leader and market leaders always do that. Managers are extremely comfortable with market leaders, programmers are familiar with it, and DBAs are easy to find. That justifies the extra cost for most corporations.
It is strange that this "ask slashdot" column lumped MSSQL in with Postgres and MySql. MSSQL is not free, and an enterprise license is not cheap (>$20k if I recall). I have very little experience with MSSQL but it is generally not well regarded.
The most important thing is to keep things portable across databases. The biggest danger with the high end database is their vendor-specific extensions keep you locked in. Once you've gone Oracle and done all the Oracleish things, it's hard to undo that. I fell into that trap with Informix.
Start small and keep it portable. Chances are that a solution like PostgreSQL will work quite well. If for some reason you start hitting a wall, it should be easy to move up if you were careful.
And that's what it's all about. Stay flexible, because it's a dynamic world.
Scalability & stability. If you don't care about those, use one of the free offerings.
We will know databases are a commodity technology when the free offerings are as good or better than the commercial offerings. Right now that isn't the case.
--
--
E_NOSIG
More to the point, if you are willing to do the work
of maintaining consistancy yourself where it counts,
you can often not only do so, but your code benefits
from it hugely. Example: my application has a few
statements which must be atomic (everybody does this,
AFAIK, certainly PostgreSQL, MySQL, MS, Ora, etc).
Now, I have one piece of code that needs to execute
a complex batch of statements with consistency and
rollback. I can easily write a library that supports
rolling back this one transaction. Oh, and look, this
one transaction also does not need to complete lightning
fast, and happens rarely. That means I can take advantage
of best speed everywhere else, and use table-level locking on this one transaction just to be safe. These are not options in most large RDBMS', but certainly are in something like MySQL. In fact, you can swap out which back-end storage system (MySQL, Berkely DB and soon Progress) to use on a table-by-table basis for best reliability/performance characteristics.
Bottom line is this: Oracle has it's place, but once you start
to need high-end performance, it falls over in the price/
performance arena. I can run the same database on
MySQL engine for 1/10 - 1/100 the cost and get the same
reliability.
What most people really need out of a database is to never
lose data. If your DB loses data, then get another
one fast. Otherwise, you should be looking at how
much you'll pay when you add the 4th processor....
It's a comples question with complex answers.
In the end you have to decide weather or not you need to scale up and when.
Of thhe free databases three that are closest to doing what oracle does are interbase, postgres and sybase 11.03 (not open source but free). Each one of these is suitable for a small to medium size company to start with. sybase is aging but still usable and postgres has a quirky set of features but keeps getting better at a very fast rate there are also a couple of companies prioviding support for it. Interbase has been around forever, is well tested and has a large userbase. It is supported by an exrtremely knowledgabel company called ibphoenix and borland too. It also runs equally well on windows or *nix and has great third party client side tools for administration.
The next tier up in price you have ms-sql, db/2, and informix, and sybase. MS-SQL is ok if you are willing to tie yourself to windows forever and are willing to pay extra for jdbc drivers and don't want to ever access from unix with php or perl or python (freetds is a valiant effort but it's still not 100% functional).
DB/2 is surprisingly affordable for the workgroup edition I was quoted about 6 grand for an unlimited user license. Compare this with the 5K per processor license for sql2000 and it's a bargain. Your mileage may vary depending on the size of your company and how afraid MS is of you switching. They offer deep discounts to large clients.
I have never used informix or sybase so maybe other people can talk about those but sybase does offer sqlanywhere studio which is a low cost alternative you can check out.
Mysql still has a way to go catch up with these databases so I would say forget about it for now.
In summary. The free databases do anywhere from 70-80% of what the big bucks databases do at about 70-80% of the speed. If you don't need that last 20% there is no need to spend the money.
Allthough none of these databases can compete with oracle head to head they would make great starting points for a company just starting out.
Of the open source databases Interbase is easiest to setup and use IMHO so it would get my final recommendation.
War is necrophilia.
There is an excellant middleware for delhi called ASTA look into it.
War is necrophilia.
In my opinion, it most definitely can be, if the business requirements are used as a primary factor in the database engine selection;
... dynamic tuning, and those dreaded vendor-specific extensions are often used in support of this area. In a modern enterprise, expect millions of transactions (each transaction could be updating 30 tables, with several hundred rows) on an hourly or daily basis.
... 3+ dimensional location data). Similarly, can I distribute my data (I'm not simply referring to tuning) so that it resides where it is required? Without thee included features, what effort must I make to program around the DB engines deficiencies (usually called 'undocumented features' by the vendor).
1. Availability: how crash-proof is the RDBMS is 1 part of the equation. Another is the ability to backup and recover the db. In a 24x7x52 shop, you virtually never get the opportunity to shut down the DB. A simple file or filesystem backup will NEVER let you recover a transaction-consistent system. Also, how well does the engine support high-availability clusters, with automatic replicatin and hot-standby servers.
2. Performance: This is the 1 area where no product is perfect, but the tier-1 products such as Oracle & DB2 particularly excel
Application availability: A company doesn't purchase a database just to have a database, anymore than they should purchase hardware or an OS for their own sake. There are more applications such as ERP and CRM, as well as many client-server and two-tier apps written to support the tier-1 engines.
Scalability: Its fine to cluster servers; I wouldn't want to tailor postgres or mysql to distribute my database on many servers in a number of different geographic areas. Tier-1 vendors have addressed this in the vendor-specific tools (although ANSI 92 attempts to incorporate many of the special features). Size is another factor; can the engine process terabytes as easily as gigabytes? If you have a server with 20GB RAM, will the database take advantage by caching objects (either automatically or by administrative instruction)?
Support: While support contracts CAN be purchased to support most anything, can you call the vendor at 8am Xmas day with an issue that has caused the database to crater? I know from personal experience that Oracle provides that level of support with reasonable expertise available.
Features: While this is a double-edged sword, can the database support all of the required corporate data entities? I include single fields of 400MB (Seismic traces and well logs), as well as providing indexing strategies that reflect the application (i.e. spacial data
Cost: The support and development cost of applications for a free engine MAY actually be higher than purchase of a packaged application for an expensive db.
Security: How does one secure transactions (not just https from the web) between the visible server and the application servers and the database servers? For example, in the petrophysical world, the most dangerous information for competitors to gain is the knowledge of the latitude and longitude that a company is querying.
There are few clear-cut answers in a general sense, but I believe that it is fair to say that the more demanding the application, the higher the probabability that a high-priced RDBMS will be required.
Bill
By itself, no, but there are a number of organizations who will support mySQL for you.
It's no different than support for any other open source software. If you want training and support, you find a third-party company that does training and support.
They may not even really be "third party". Randal Schwartz' Stonehenge Consulting does Perl training. Randal is a key developer in the Perl community. Does he count as third party?
--
Also keep in mind that Sybase offers Linux x86 binaries of it's previous major release for free for any use (deployed commercial use okay!). They offer the current version free for development purposes. The previous major version is plenty good enough for most applications. There is no comparison between MySQL and a real industrial strength database like Sybase. Maybe one day MySQL will be up to snuff and then we'll all use it, but Sybase and Oracle have been working on this stuff for a long time.
Burris
Along those lines, there's also GemStone/S and GemStone/J, OODBMSs for Smalltalk and Java respectively. I don't know about the Java version, but the Smalltalk version is pretty cool stuff, and can interface with C and C++. A lot of people prefer not to call it an OODBMS, but an application server. It's both, more or less.
Working toward a usable PDA environment in the spirit of Newton OS: Dynapad
Good comments/theads pertaining to this were...
Why do we always have to see stories in groups of 3 (that means there's one more DB story coming tomorrow)?
I just came across this, and it seems to address the topic quite frankly. It is, however, wuite simplistic and provides little technical detail. Perhaps someone more familiar with the field could comment...
Your information is old; MySQL does have transaction support now. http://www.mysql.com/news/article-18.html
Yet another RDBM debate cooking up! This is like asking whether there is any point to mission critical software since the mainstream works just fine for the majority of cases. See here lies the point the majority doesn't need an Oracle database just like the majority won't ever need high end servers because their requirements and constraints can be accomplished with the use of off the shelf hardware software. Specialised software fills specific niches and becasue the market for it is relatively small the sales margins have to be high to justify writing said software. If you think that Oracle licensing is steep I encourage you to look at pricetags for geoscience software products. We're talking an excess of $50000US per SEAT. Are some of the technologies more suited to different tasks? No. Not at all. EVERYTHING can be shoehorned to run on a pentuim box with Lenux preinstalled on it. You can be sceptical all you want but once you are responsible for a system where a downtime could mean lives lost or millions wasted you will realize that a meager few thousand bucks is just peanuts. How much would you be willing to spend to protect the life of your kid? Or your wife? Or your sister? (think air traffic control software etc.) Please think those things through before posting them to slashdot front page.
Your pizza just the way you ought to have it.
Just like that old phrase says, noone ever got fired for buying Microsoft. The same goes for Oracle.
There is also the faction that will say that MSSQL cannot scale up to what Oracle can. I tend to agree with this line of thinking, but to each his own. In the end, most companies that have the money will end up purchasing something like oracle, just because they know they have a billion dollar company that has proven itself again and again and again.
Just my .02USD
When I first started working with Oracle (I'm a certified DBA now), I often had the same skepticism. I had worked with MySQL before, and couldn't understand what all the hype and the expense was for. What's the big deal about storing data in tables and querying them with SQL statements? No doubt there's a bit of sophisticated computer science that goes into that, but surely there's a limit to how complex it all can get. Why the hell does Oracle have to be so damn complicated (I hated trying to understand the docs, which always leave out the forest for all the trees), not to mention expensive?
After a while, I understood that as far as representing tables and data and SQL statements is concerned, this really is true. About this part, the questioner is right: An open source solution will represent your data just fine, no reason to pay out the nose if that's all you need. And MySQL is still almost always faster than the expensive commercial RDBMS's. But tables and SQL statements are only one small part of a database application, and may not be the most important part.
What you're getting for all that complexity and cost is all of the extra services that come with the data management, and these things really are worth it in many projects: Constraints that enforce data integrity, safe transactions according to the ACID principles, robust backup and recovery regimes that really can survive the worst catastrophe, distributed computing and clustering (for high availability and distributed load), and a lot of parameters to tune. All of this stuff seems like add-ons, if all you think about are tables and SQL statements. But for many applications, the data model is not what matters most; these things are by far the most important qualities of your RDBMS.
Oracle RDBMS's still could stand to be less expensive and complex; but I see the justification now much better than I used to.
Always keep a sapphire in your mind
... you've already answered your own question. When you need Oracle or DB2, believe me, you will KNOW it, or you're a moron and beyond hope anyway. At my current position I'm managing a Web team responsible for a site that lets users build complex queries that then sling around 16GB databases. We're using MS SQL Server and even with that we look covetously upon Oracle's reliability and scalability. There is no way on God's green earth I'd try to do this in MySQL. I love MySQL, but it's just not ready for prime time for these kinds of applications. Which is why I say you'll know when you need it; if there is any question in your mind, you probably either don't need it or aren't experienced enough to know what you need in the first place, so be careful that you don't put your Free Software Advocate hat on over your Responsible IT Professional hat :-)
-- Jason A. Lefkowitz
Read my blog.
Then there's the case of the company who's DB is worth $1M/hour when it's down. Who are you going to sue if MySql goes down for a day? Disclaimers aside, Oracle actually has some money you can go after.
Where the development has already occurred on an OS DB, there's lots of reason to keep it there (as long as it doesn't break because of an inherent design problem/bug). Similarly with a big $$ system. In either case, the cost of the database is usually small compared to the cost of redevelopment and/or breakage.
--
Free Software: Like love, it grows best when given away.
Although I love and use MySQL (with PHP and Nix), it is DEFINATELY NOT a full features RDBMS similar to Oracle. (last I saw, it didnt support Views, sub-selects, and a buncha other stuff). Interestingly enough, we are starting to use Oracle 8i here are work, and it offers some neat things like running Java Inside the DB (which makes data-processing faster, but puts the load on the DB Server). Also Oracle instance replication is neat.. you can just have a buncha boxes running Oracle with something similar to Hot-Swapable Databases (think Load Balancing, which I know the MySQL ppl where playing with at one time).. I'm no Oracle Expert, but it is definately More robust then MySQL / Access / MiniSQL, etc, but I would only consider it for Enterprise level apps.. -- Insert whitty line here
----- The internet has given everyone the ability to have their voice heard equally as loud.. even if they shouldn't be
IMHO, Oracle is not the price/performance leader in the commercial RDBMS market, nor is it the easiest with which to develop or to administer.
Sybase has long combined relative ease of use, development, and administration with pretty good performance. Lately, though, it seems Sybase would rather be in a different market -- not sure what market, though.
For projects which require a commercial RDBMS, I now choose DB2 UDB v7.1. And, yes, I tend to run it on Linux. It falls between Sybase and Oracle in the ease of use, development, and administration department. Performance-wise, for the kind of work I do, it's at least as good. Price-wise, it's a better deal.
Culturally, I rank Oracle alongside Microsoft. No, thanks!
PostgreSQL is under active and aggressive development, which is bringing major new features, such as outer joins, to their product. Version 7.1 should be out by mid-March. For small and medium projects with limited budgets, it's almost ideal. (I'd like to see stored procedures which can return results sets.)
There are many smaller companies and organizations who can benefit from web-based applications, but who have limited budgets. All of the commercial RDBMS products have fairly hefty licensing for Internet connectivity. That can price many of these organizations right off of the web (or force them to violate the license agreement, exposing them to costly litigation). PostgreSQL (and MySQL, I'm sure), are a viable solution for these businesses.
When I get some time, I'm going to try implementing one of my DB2 projects in PostgreSQL 7.1, just to see how it does. I already know I have to re-write the stored procedures as queries, so there will be a labor impact. It will be an interesting experiment.
Gordon.
He that breaks a thing to find out what it is has left the path of wisdom.
-- J.R.R. Tolkien
it's quite informative (at least read the first chapter): SQL for Web Nerds by Philip Greenspun of Philip and Alex's Guide to Web Publishing fame.
Seriously, go read it. It's an afternoon well spent, in that hopefully after reading it you'll understand why comparing MySQL to Oracle is comedic. You have to understand your tools before you pick which one to use (MySQL is for your MP3 id3 tags, Oracle is for your hospital's data storage or for your credit card company's data storage.)
--
News for geeks in Austin: www.geekaustin.org
News for Geeks in Austin, TX
North Carolina is entrusting their medicare system to some dude writing one page porn sites, are they?
What do you pay for? Well, mostly you're paying to see your pillow more often, or at least that's what the vendor is trying to sell you -- peace of mind.
:-).
I have (lately -- SQL 7, not just SQL 6.5) seen MS SQL databases that buy the farm every night and get rebooted, and this is referred to as "more stable than before". I've watched MySQL crumble under heavy concurrent read/write loads on large tables. I still haven't figured out an easy way to carve up PostgreSQL backend storage so it scales more like Oracle. And I've thus far managed to make every (yes, EVERY) database I've ever used core dump. I'm working on DB2 at the moment; apparently it runs across x86 Linux clusters, which of course is something I'd like to crash. I got Oracle 8i to crash (finally!) earlier this summer. That was unusually difficult; Sybase and Informix were relatively easier. So, when you go the commercial route, you're paying for features, branding, and a lot of stuff you already get from the free RDBMS packages. Plus, more people tend to know about the bugs (every last RDBMS I have seen has at least some bugs).
BUT when you get right down to it, even Postgresql (my favorite free database) simply does not have the application support or PROVEN OLTP scalability of something like Oracle. Today someone batted around the idea that "DB2 is okay up to around 10 or 15TB, but I talked to some banks and oil companies, and they say that beyond that, it just falls apart". It may be the case that Oracle scales a little better than DB2, but up until recently, the top 10 (size-wise) OLTP databases on earth were running NCR Teradata, IBM DB2, and I think there may have been one Oracle customer in there. UPS's database, the largest active OLTP database, was only around 12TB last time I looked.
12TB of live data is a HUGE amount. EMC's top-of-the-line Symmetrix, a piece of equipment costing a cool $3.2M, just barely fits 12TB if you're running everything mirrored. God forbid you blow out the cache on that monster. And of course there's the question of backups...
Basically, if you're tied into a vendor's specs (eg. "We only support MS-SQL here at Affymetrix" or "ArsDigita only supports Oracle as our ACS backend") then you don't have to choose, just get out the wallet and bend over. No lube for you!
Otherwise, in my neither limited nor hugely abundant experience, MySQL and Postgres work pretty damn well for read-only data marts and low- to mid-volume read-write backends.
Beyond that -- if MySQL can't handle your OLAP needs, and Postgresql can't handle your OLTP needs -- then maybe you have to look at Oracle or DB2 or whatever. But by then I strongly suspect you'll know why they charge so %$@#ing much
The differences are fairly marginal, but if you need them, you need them, and that's the end of it.
Remember that what's inside of you doesn't matter because nobody can see it.
Like it or not, Oracle is worth every penny, assuming you can afford it. If you, or your client, cannot afford the (admittedly) big bucks for an Oracle licencse, there are plenty of free (and Free) solutions available.
What do you get with a 10-30K (or, in my experience, a 50-80K) license? You get Oracle's years of development and expertise in managaging some of the biggest and baddest installations. Think 500-1000 gigabyte DBs in constant use for 15 years.
You get a whole bucketful of trained and talented DBAs (who are worth every damn penny of there high 5- and 6-figure incomes) who grok Oracle syntax.
You get Oracle Corporation's culture (at least, I hope it's still the culture). At Oracle, the technologists create cool stuff and pay the marketing department to find or create a market for said cool stuff, rather than marketing dictating what technologies are created. Thus, when you visit Oracle, you find 3 or 4 products that do nearly similar things, or overlapping technologies. (I think this is a good thing -- others may disagree)
Finally, you get name recognition. Don't laugh this off -- it's worth a lot of money to some people. If I say "Let's do it in Oracle", people respond, "Yeah, that would be best." If I say "Let's do it in PostgreSQL", I hear "Huh?".
It boils down to: If the money is there, spend the money on Oracle, and advertise that fact. If the money isn't there, use PostgreSQL, and just say you have a RDBMS back-end.
(Oh, MySQL isn't a RDBMS in my humble opinion. It's a fine solution for certain problems, but don't confuse a SQL interface to a filesystem with a true ACID-compliant RDBMS. Future versions of MySQL promise to alter this perception, which may change the dynamics.)
Potato chips are a by-yourself food.
There are a number of reasons why you would pick any particular RDBMS, but I have seen three which come up most often and are of the most importance.
Support
This is a big one. And an enterprise scale RDBMS needs a lot of support. Ever wonder why some Oracle DBA's make more than skilled developers? Administration of a large database is not trivial, and these databases often run the core of a corporation. Mission critical to the extreme is common for these systems (Hence the reason Oracle can charge millions for a large Oracle Applications ERP installation)
Performace
These big databases are built to handle huge data sets. Oracle supports all kinds of direct interfaces to read/write directly from a SCSI/Fibre channel RAID setup. A number of the Oracle extensions, or custom application also have to be extremely fast given the amount of data they are working. PL/SQL and embeded SQL, OCI, etc, are all tools to allow maximum performace with large heaps of tables and rows.
Add-ons
And by this I mean extra pieces of functionality or applications that are only provided with a specific RDBMS. Oracle Applications, which is a popular ERP package, requires an Oracle database. Likewise, PL/SQL is very popular, but only available for Oracle. Add in all the other niceities that Oracle ships with 8i, and you can see why the entire package is pretty impressive.
So, in the majority of situation a smaller opensource RDBMS like postgreSQL or mSQL will probably work just fine.
If you are n enterprise customer which needs extreme scalability and performace, You would be hard pressed to go with anything other than Oracle or IBM.
This all sounds simple enough, but producing a db that actually implements the ACID properties reliably is very, very hard. So hard that people tend to charge tens of thousands of dollars for the result.
Of course, in many circumstances you don't actually need a fully ACID-compliant db. In general, if your transactions don't involve strangers' money, and if data can be cross-checked and reacquired from external sources in the event of a problem, you can get by with a cheaper, non-ACID db server. But using a non-ACID db where you really need ACID-level reliability is asking for big-time trouble.
--
When all you have is a hammer, everything looks like a skull.
In particular, built-in enforcement of data constraints, triggers on data modification, and stored procedures are HUGE wins for many web sites. Using these correctly can move tremendous amounts of complication out of your PHP or ASP code and into the database where they can be more easily tracked and managed.
The lack of transactions in MySQL is often commented upon, and its very important. A complex set of related changes can fail in the middle for many reasons, leaving your data in an unknown or even a completely corrupted state.
Is it worth all of that to pay Oracle's fees? For a commercial web site, where dollars are at stake for every bit of downtime or lost data, then the answer is absolutely yes.
If, on the other hand, you're working on something that uses the database as structured storage rather than a full-fledged DBMS, and especially if you don't mind babysitting it from time to time, the cheap or free options are probably OK.
Anyone who has ever been a DB developer (such as myself) or a DBA knows tha thte vast majority of enterprise-class databases (Oracle, DB2) are NOT used for web sites. Web sites are cute, but rarely are they mission critical, and rarely do they hold vast amounts of information. Think large systems. Cam you imagine, say, Visa or Matercard running MySQL to handle thousands of transactions a second? Or how about truly large databases? I recently set up North Carolina's Mediare system on Oracle. That started out as a 3 TB database, and will grow steadily for the forseeable future. Has MySQL ever been tested against something of that size? Even the largest web sites do NOT generally compare to other entreprise-class projects, in either size, ort he critcality of data.