Oracle To Offer A Free Database

An anonymous reader writes "ZDNet News reports that Oracle is likely to announce a free version of its Oracle 10g Database. Oracle Database 10g Express Edition will be free for development and production use, and could even be distributed with other products. What does this mean for the future of MySQL and PostgreSQL?" From the article: "By introducing a free entry-level product, Oracle intends to get more developers and students familiar with its namesake database, Mendelsohn said. Those customers, Oracle hopes, will eventually upgrade to a higher-end version."

Comments

First Post

[CptChipJew]

I can't wait to get this. I really enjoy using 10g for projects, and hopefully I can distribute this with my free software.

My one question is, how will this stack up against MySQL. Other than the fact that MySQL is supposedly more free, how could it beat this?

Propz to GNAA

Re:First Post

[$cullyshouse]

At the risk of starting a flame war....

... Oracle is a rubbish dinosaur that hasnt aged all that well. DB/400 (and its ancestors) is older but better why? coz IBM said if it dont do what you want, work round to it. Oracle said, ok we'll patch it. MySQL is excellent for what it is, a website database server so i dont think they have to worry too much I cant see many php developers going to the trouble of using oracle.

Re:First Post

[badfish99]

If MySQL will do what you want, then you don't need Oracle.

But if your database is really big enough to need Oracle, then MySQL certainly won't be in the running as an alternative.

Re:First Post

[Short Circuit]

If FreeDOS will do what you want, then you don't need Windows or Linux.

But if your application needs to support enough different network drivers, then FreeDOS certainly won't be in the running as an alternative.

It's mostly about branding and attracting the attention of PHBs. There's probably some interest in scalability, too.

Re:First Post

[Zathrus]

Oracle is a rubbish dinosaur that hasnt aged all that well

And this is insightful?

It's a baseless accusation. The poster doesn't even attempt to provide any proof for it. Oracle is continuously leads the pack in benchmarks, it has more features than you can shake a stick at, is incredibly stable, and has features that MySQL is just starting to catch up with (wow, MySQL finally got views! How wonderfully 1980s.)

coz IBM said if it dont do what you want, work round to it. Oracle said, ok we'll patch it.

So suddenly not adding features and refusing to respond to your userbase is a good thing? No wonder IBM's lost most of the market outside of mainframes and minis.

MySQL is excellent for what it is, a website database server

Well this much is true at least. But I still wouldn't use it much beyond a toy website. PostgreSQL or Firebird are better for the same price -- both in features and in stability/reliability.

cant see many php developers going to the trouble of using oracle

The trouble? You clearly don't know what you're talking about now. Oracle is far easier/better to write SQL for since it's both more flexible and closer to the SQL "standard" (and that's a pretty sad statement). There's also far more information out there for help with Oracle than there is with MySQL, not to mention that Oracle is something very useful to put on your resume/CV -- MySQL isn't totally unknown anymore, but Oracle is still better as far as that goes.

Now if you want to rightfully bash Oracle then talk about their miserable installer and bundled administration tools. They suck. They've always sucked. And they're not getting better IMO. Oracle's on a buying spree right now, and I so wish that they'd buy out Quest Software and bundle TOAD (Windows) or tORA (*nix) with their servers. The Java crap they use now blows. The other (and related) issue is that administering an Oracle server can be a daunting task, and there's not a great deal of (free) literature available for it. Oragle 10g has made strides here with the database doing a lot of self-fixing and tuning, but it could be better (or at least better documented). Of course, one reason that MySQL doesn't need as much here is because there simply as much that can be done to it. Flexibility has a price.

Re:First Post

[kpharmer]

> If MySQL will do what you want, then you don't need Oracle.

Not necessarily true, the idea of 'just use the best tool for the job' is very myopic. For exammple, in a 500-person organization with say, twenty database servers, has a need for consistency. Why? Because otherwise they waste a lot of time:
    - training dbas and developers on sql extentions and limitations
    - training dbas on multiple database backup and restore methods, issues, and management
    - managing multiple license types
    - training dbas on performance tuning on multiple databases
    - etc, etc

In fact, you can often save money this way while paying surprisingly high database licensing costs - through reduced administrative and development labor costs. Don't get me wrong, I still like site-licensed or free databases since it makes architecting solutions so much simpler: you can create a new database without waiting months for funding & procurement. And I'm not really a fan of Oracle.

Re:First Post

[dbucowboy]

So I guess Google isn't big enough to need oracle yet? They're still using MySql for everything...

Re:First Post

[Anonymous Coward]

The Oracle Documentation Suite is freely downloadable from their own website.
http://www.oracle.com/technology/documentation/ind ex.html

Re:First Post

[Eivind Eklund]

I want the database to
(A) return the same data on SELECT as I inserted in INSERT, unless I got an error
(B) return updated data after an UPDATE, unless I got an error.

MySQL doesn't fit either of these requirements - the first by design, the second due to bugs. I'm ashamed of the use at work. We just haven't had time to migrate away :-(

Eivind.

Re:First Post

[kcelery]

it is interesting to find that no one is suggesting 'a beowulf cluster of free Oracles'.

Re:First Post

[camt]

The other (and related) issue is that administering an Oracle server can be a daunting task, and there's not a great deal of (free) literature available for it.

Daunting, yes. But on the contrary, I find Oracle to be one of the best documented databases on the market. There is an entire library of documentation you can download for free, search online, or get on CD with each release. In addition to that, there are forums on Metalink for paying customers, and the invaluable http://asktom.oracle.com/ website (free)! Most of the documentation includes relevant examples in addition to just technical reference.

Rarely have I been confronted with an Oracle issue that is not easily resolved by a combination of those reasources.

I agree about their Java tools. Anyone who uses Oracle should have some Quest tools laying around...

Re:First Post

[Anonymous Coward]

Now if you want to rightfully bash Oracle then talk about their miserable installer and bundled administration tools. They suck.

Heh. I haven't used Oracle in over 5 years and my main memory is of the astonishingly crappy Java install and the lack of GUI tools. It's incredible they haven't fixed it.

Re:First Post

[Eunuchswear]

Uh, got some kind of cite for this?

Re:First Post

[dbucowboy]

Nope... just common knowledge... if you find anything disproving me please let me know.

Re:First Post

[Anonymous Coward]

ya, i think google is using it as part of bugzilla, gforge, etc, etc.

is it using it for its search engine? ah, nope.

Re:First Post

[Anonymous Coward]

Even small databases may not be suited for mysql, if you want reliability.

Re:First Post

[Anonymous Coward]

Their financial data provider uses Oracle.

Re:First Post

[Anonymous Coward]

Too bad TOAD also seems to be full of bugs. And yes, that java crap is Mostly Useless (tm). Also concerning PL/SQL, I don't know what that language was designed for. An almost-procedural language for controlling an almost-functional like sql?

Sure, why not... if you can make it work. Too Oracle isn't one of those cases.
A little bit of syntactic sugar here and there and removal/deprecation of obsolete features could do wonders here.

Now that I've got those things said, anyone have real-world experience of Ruby on Rails with Oracle? Or the python equivalents, etc? :-)

Re:First Post

[Anonymous Coward]

No wonder IBM's lost most of the market outside of mainframes and minis.

These market share charts say otherwise:
Application Server http://www.sqlsummit.com/Trends/AppServerMarket.ht m
Database http://www.sqlsummit.com/Trends/DBMSMarket.htm

IBM had $96 billion revenue in $2004.

Re:First Post

[zardo]

You obviously have no idea what you're talking about.

Re:First Post

[badfish99]

I bow to your superior knowledge. You have obviously been involved with more enterprise-class database systems than I have.

Re:First Post

[Anonymous Coward]

Anyone doing so would not have read the license. There is a very specific market for this Express Edition. But it would be pretty typical for bitheads to ignore licensing and requirements, just 'cause.

Re:First Post

[Anonymous Coward]

You can make MySQL scale if you have to, and it can be a viable alternative if you are willing to engineer around its scalability issues. Sabre Holdings currently have a 2 Billion row deployment and the largest I know of, is Los Alamos, with 7 Billion rows. I'm currently working on a 8-16 Billion row MySQL platform which will hold around 4.5TB of data.

what a wimpy database

[defMan]

Database XE is free for runtime usage with the following limitations:

• Supports up to 4GB of user data (in addition to Oracle system data)
• Single instance only of Oracle Database XE on any server
• Only uses and executes on one processor in any server
• Can use up to 1GB RAM

Re:what a wimpy database

[popeyethesailor]

Think embedded DB.. This is just for applications which need a local cache. Dunno how much real muscle this has inherited though..

MS offered a SQL Server express too. This is probably just a response.

Re:what a wimpy database

[mysticwhiskey]

I think that for embedded database storage, this would be overkill.

Re:what a wimpy database

[m4dm4n]

While the summary asks the question how this will effect MySQL and PostgrSQL, surely the limitations on processors, memory, and instances will make this unusable for shared hosting. While I may be wrong, I bet a lot of people get to know MySQL and PostgreSQL when they get it as part of a package for cheap (and thus almost certainly shared) hosting.

Re:what a wimpy database

[Professor_UNIX]

This seems like it's aimed at Microsoft's "free" MSDE rather than open source databases like Postgres or MySQL. The specs are on par with MSDE.

Re:what a wimpy database

[defMan]

Aha, i didn't know this. That explains the specs a bit more.

I don't see it as competition to opensource offerings, because a) it's not opensource and b) it's extremely limited. It's main use is to install it on a developer machine to make sure they don't mess with the real database.

And anyone considering this for embedded should probably go for sqlite instead.

Re:what a wimpy database

[smittyoneeach]

The thing about PostGreSQL that trumps the competition, IMHO, is that you can build in support for tools such as Python on the DB server.

PL/pgSQL bears a resemblance to PL/SQL, and both languages are servicable enough. Oracle cooks in its own JVM. While Java is an undeniably powerful tool, one feels relatively enslaved to the JVM, compared to the bliss of simple, clear Python code.

Re:what a wimpy database

[Goo.cc]

I believe that these limitations mirrors Microsoft's SQL Server 2005 Express Edition, which is where they probably got them from.

Re:what a wimpy database

[ShootThemLater]

I think that for embedded database storage, this would be overkill

Not necessarily. Some vendors offer products that use a database to (say) store metadata. It's not unusual for such vendors to only offer support for databases where there is sufficient proven commercial demand - and (right or wrong) in many fields, that means Oracle, DB2, MSSQL. Sure, something like MySQL would be far better suited for embedded use, but that would be a whole other platform to do QC on etc. and it might not make commercial sense.

So, if you had a free version of a database that you already support, you could easily use that as an embedded version - for a mobile version of your product, for example.

Re:what a wimpy database

[mysticwhiskey]

I'm just thinking of the overhead of Oracle, which may not be that suited to an embedded application. Of course, if this free version has lower hardware requirements, then yes that would be appropriate. However in this case I doubt that this would be suitable for say, mobile devices.

Re:what a wimpy database

[Anonymous Coward]

Yup, totally.

I can use it in a non-production environment. But once it's going into a production environment, it's gonna need more than "4GB of user data". What's the point of XE anyway? If it's only for trial purpose, I can just download a full version and use it in a non-production environment.

I'm gonna stick with MySQL

Re:what a wimpy database

[way2trivial]

except microsoft counts dual core intel chips as one processor for purposes of licensing..

doesn't oracle count cores?

Re:what a wimpy database

[LLuthor]

PostgreSQL also can embed a JVM for writing stored procudures and user functions and aggregate functions, but its not very well supported (yet). PG does have quite a few companies behind it so I doubt it will stay that way for long.

PG probably has the best language support of all DBs. Is there any major language that doesn't have a PG interface in 8.1?

Re:what a wimpy database

[craigmarshall]

How do these limitations stack up against something like PostgreSQL or Mysql or anything else free as in speech? i.e. Can they take advantage of multiple CPU and are there limitations on the database size? Any RAM limits for whatever reason? Just interested.

Craig

Re:what a wimpy database

[ShootThemLater]

I'm just thinking of the overhead of Oracle, which may not be that suited to an embedded application. Of course, if this free version has lower hardware requirements, then yes that would be appropriate. However in this case I doubt that this would be suitable for say, mobile devices.

I agree, it has a significant overhead - particularly in memory and disk footprint. I should clarify though - when I referred to mobile use, I was thinking of users with reasonably high end laptops who need to use a database-dependent application while offline. So while in the office, they could connect to a main server running 'proper' Oracle with live data, and when on the road, they could use a personal installation with the XE server and cached data. The advantage for the vendor is that they can use the same database for both.

Re:what a wimpy database

[Dan_Bercell]

MS has always offered a free database, MSDE

Re:what a wimpy database

[strudeau]

Most LAMP apps and sites live in hosted environment. For many large scale hosts, this licensing will not be attractive, so they'll continue to offer MySQL (and sometimes PostgreSQL). And since that's what's available cheap-and-easy, I suspect the open source databases don't have much to fear.

Re:what a wimpy database

[cduffy]

This is just for applications which need a local cache.

Yes, but Oracle *sucks* for that. SQLite is considerably more appropriate for that kind of usage -- it's much less of a PITA to administer (and Oracle *is* indeed a PITA).

I'm quite certain this is aimed where they say it is: Open-source databases such as PostgreSQL and MySQL. Many of the instances of either are small installations intended to act as backend to a webapp or five for a small company's intranet -- nothing needing hardware outside of the range of what they're limiting this edition to.

That said -- there are cases where an Oracle database can be hosed badly enough that the recovery docs specify to always call Oracle support before going through the next steps provided. Support's not going to be free even if the software is, right? And Oracle gets pretty darned arcane.

Heck, this'll give more folks a chance to see just how bad Oracle is about not complying with operating systems' local conventions (FHS? What's that?) and requiring vastly more specialized knowledge to admin than their OSS competitors... perhaps it'll even *lose* them marketshare.

Re:what a wimpy database

[LurkerXXX]

Mod parent up. I've seen lots of posts on /. about how SQL Server Express is some new idea. MS has been giving away a lightweight version free for a long time.

Re:what a wimpy database

[oldwarrior]

I think it is aimed at open source free databases like mySQL. After getting embedded into developer projects, Oracle will try to entice them into getting the "professional" version with a long list of tasty plugins/modules not available to the free one. It will work, too.

Re:what a wimpy database

[EraserMouseMan]

You should RTFA. The strategy is directly aimed at gaining more usage from the people who typically would choose MySQL or PostressSql.

Microsoft's SqlServer 2005 express has the same strategy. But Oracle is doing this for the same reason Microsoft is --> they are getting jealous of the Open Source database market share.

My personal prediction is that Oracle's lite version won't catch on because Oracle's db is so dang complicated to set up correctly and the tools stink in comparison.

Re:what a wimpy database

[photon317]

Sounds like competition for MySQL then :)

Re:what a wimpy database

[mikaelhg]

I wish that Microsoft will up the ante and remove the memory and size limitations!

Re:what a wimpy database

[Jeppe Salvesen]

So what? If it comes with the required features, I'll still be able to make sure my product works with Oracle - without paying Oracle a dime. In my view, that's perfectly reasonable - since my customers will buy the expensive Oracle versions..

Re:what a wimpy database

[Anonymous Coward]

Oracle has offered a database specifically aimed at mobile use for quite a while now. It's called Oracle Database Lite Edition. XE is a different animal more useful for the person who wants a quick-to-install, easy-to-administer Oracle database to use as a learning tool, a development platform, or a useable database engine to include with a trial copy of software that they've written to work in an Oracle environment.

Re:what a wimpy database

[ll1234]

Each core counts as 3/4 a processor. Oracle had wanted to count each core as 1 processor but had to back down in the face of customer backlash and the Microsoft/IBM decision to count each physical die as a processor.

Re:what a wimpy database

[dindi]

4gb of data : if that is the max it can use I still better go with mysql as I found 4GB to be the problem point with locking and all that on the hardware I could afford ....
besided the only prject I would use it is kicking it's 12th GB so I can't even try oracle XE there

Re:what a wimpy database

[bataras]

I have used MSDE in development (for production sql server deploymnet) as an actual production instance for a small website and in an embedded situation (a settop box). MSDE is SQL server stripped down much like this free version of Oracle. Two super important differences between MSDE and Orcale Express (if the Oracle Express description above is accurate): a) MSDE only runs on windows b) MSDE has a connection governor in addition to the other limitations. This "governor" will ON PURPOSE start slowing down queries when you get to like 8 simultaneous queries.

Re:what a wimpy database

[blowdart]

Each core counts as 3/4 a processor

True. But the processor count is rounded UP. So if you want to license a single dual core machine, it's 1.5 licenses, rounded up you must purchases 2 CPU licenses.

Re:what a wimpy database

[tetranz]

The replacement for MSDE, SQL Express does not have a governor. Its limited to a single CPU, 1GB RAM, 4GB maximum db size so, as others have said, the free Oracle product looks to be an answer to this.

SQL Express is pretty good for many small websites, especially for DotNetNuke sites or other ASP.NET stuff.

Re:what a wimpy database

[Eivind Eklund]

MySQL isn't suited to anything where you need a database that actually works. No, not with InnoDB either - it e.g. corrupts data when you change the record in the primary key index. Yes, I work with MySQL, and I've worked with PostgreSQL (the reason I don't do so now is perceived migration cost.)

Eivind.

Re:what a wimpy database

[bataras]

Vewy intwesting. Yeah, here's an MSDE - Express comparison:

http://msdn2.microsoft.com/en-us/library/ms165672. aspx

BTW, "embedded" a DB like this is not goot for memory. MSDE eats up 10's of megs of memory over time with just a simple schema. I've seen it at 200mb mem usage for a DB with a 2mb dataset.

For embedded DB work, people should take a look at Firebird.

Re:what a wimpy database

[Anonymous Coward]

You can use any language that can produce JVM bytecode.

Re:what a wimpy database

[cloudmaster]

You've been able to use Oracle 10.whatever for free for development use for a while now. I took a DB class a couple of semesters ago (one more stupid requirement for an experienced sysadmin - 10+ years - to get that damned piece of paper some short-sighted employers are so fixated upon) which use Oracle and eventually added some Cold Fusion. My two least favorite web/database products. Anyway, I grabbed what may have been Oracle 10g, and installed it just fine, for free, from their website. I guess I could reasonably be using it as a backend to my web stuff right now - it's not like the outside world has any way of knowing what DB is providing data to my PHP stuff (not to Oracle licensing nazis - I'm using MySQL, actually, but just making a point). So, the main use is probably to get that already freely available development platrom into some production use, so the potential users can see hwo well it really works in their production systems - eventually buying "the big version" when they need more power. Or, possilbly getting Oracle so ingrained that it's what they wanna use when they go elsewhere - kinda like Microsoft giving software to highschools and colleges...

BTW, BlueDragon CF server is a nice free Cold Fusion development platform, if you like that kind of thing or need it for a class... :)

Re:what a wimpy database

[Unordained]

To clarify: Firebird has an embedded version containing the full Firebird database engine (full sql support, etc.) as a .dll file. The main limitation of the embedded version vs. others is that it takes exclusive control of the database file(s) it attaches to, so only one Firebird-embedded application (at a time) can be reading/writing a particular database file. The advantage is that the .dll has all the appearances of a bloated remote client library -- in fact, you rename the .dll to the name the application would have expected for the client library. An otherwise client/server application can be switched back and forth between embedded and client/server by simply swapping out the .dll it uses. You can develop your application with client/server, multi-user environments in mind, but then make your application available to home users who don't want a Firebird server process running on their machines 24/7, with a port open for connections. Stinkin' easy to do. And you don't have to give up good SQL and database integrity support to do this.

As I recall though, the embedded .dll is only available under Windows (no .so equivalent?) I'm not sure why. Maybe they figure that under linux, you'd just run the full server process?

(Firebird, like Postgres, runs on many platforms. And it's open-source. And it's free. And it's accessible from a variety of languages. And it has all sorts of neat SQL support, is extensible, is ACID, has had MVCC from the beginning, has SP's, triggers, check constraints, hot backup, etc. In other words, it's that other-other open-source DBMS every forgets about. It just happens to also have an embedded version.)

Re:what a wimpy database

[T-Ranger]

Only 1GB of RAM? Last time I tried out the demo version of Oracle, 1GB of ram was about its minimum requirement!

Re:what a wimpy database

[bataras]

What he said. Yes Firebird is pretty cool. Although on occassion we refer to it as Fireturd when trying to figure out some nuance in it's configuration/use. All DBs have those issues, but only Firebird lends nicely to a mocking name like Fireturd :)

Re:what a wimpy database

[mpeppler]

Essentially the same specs as Sybase's ASE 12.5 / 15.0 Express Edition, with the difference that Sybase doesn't stop you from running multiple instances on the same machine.

Re:what a wimpy database

[HiThere]

I wish that Microsoft will up the ante and increase the memory and size limitations!

I don't think they're stupid enough to do so, but what does reality have to do with wishing. I also wish that MS would just dry up and blow away, and take the other EULAs with it. (Not the companies using the EULAs, just the EULAs. Few companies are so bad that I wish they would just disappear. MS, however, qualifies.)

Re:what a wimpy database

[electroniceric]

I tried using Firebird for a 6-client project 2 years ago, because Postgres hadn't come out on Windows at that time. I found the documentation pretty weak overall, the admin tools were very immature, and the completeness of the features was overhyped. For example, I never could use updatable views with Access (they work just fine with Postgres), and never could nail down acknowledgement from anyone that it was a bug. Maybe things have gotten better since then, but I was totally underwhelmed. I switched to Postgres and have never looked back.

Re:what a wimpy database

[Anonymous Coward]

Good God, another brainwashed Python dipshit. Python sucks ass, you n00b.

Re:what a wimpy database

[booch]

I think you may have discovered the key target market. This will allow Open Source and other Linux-ish applications to support MySQL, PostgreSQL, and Oracle out of the box. This will give the users the ability to choose (perhaps the pay version of) Oracle for their implementation.

Re:what a wimpy database

[justasecond]

Insightful my ass. Oracle has always had a developer edition. Free. As a developer you're getting nothing new from this.

Re:what a wimpy database

[Unordained]

I'm glad you found a solution that meets your needs and preferences. While I do what I can to bring up Firebird in discussions of F/OSS DBMS's, it's mostly to give people more choices. Our project has been using Firebird for several years (somewhere around 50 concurrent users on average) but I'm quite impressed by Postgres.

I won't be able to answer your concerns specifically, without a little elaboration. But the documentation hasn't improved much; I hear it's being worked on, but I've not seen a replacement for the Interbase 6 documentation we still use on occasion. There are several new admin tools available, but we don't use any of them. We stick with isql and ibconsole most of the time. (We've built our own, into our project, with some features only we would care about.) Updatable views are a problem, in the sense that Firebird (nor, that I've seen, any DBMS) doesn't support them in every conceivable case, but there is automatic support for simple cases, and you can write triggers for complicated (but technically logically sound) ones. I don't know about using them from Access though -- that sounds more like a driver issue. As we don't use any of those drivers, I haven't kept up on their status.

Again, I'm happy you found a solution that suits you. I encourage you to keep Firebird in mind as an option for future projects, at least as something to look at and ignore. I always prefer to have some throw-away options, like MySQL. Makes me feel free.

Re:what a wimpy database

[zeke2.0]

The windows version of xe supports .net and clr, so just about any language that can compile into clr can be used, python for .net, delphi.net, vb.net,c#..... or you can still use the oracle call interface (c++, delphi) and stick with Java if you like. The java interface is very clean and easy and a viable replacement for PL/SQL based packages. BTW, I'm just agreeing with and expanding on your point

Re:what a wimpy database

[NutscrapeSucks]

Well, it hasn't always been free ... Only "Free with purchase of SQL Server CAL", or "Free with purchase of Office Professional". In fact, I'm still not sure if its totally free for use.

Re:what a wimpy database

[hey!]

MS has always offered a free database, MSDE

Yeah, and worth every nickle. I could puke a better SQL implementation than anything MS has ever created. MSSQL is likely the most non-orthagonal language I've ever worked with, and buggy as hell to boot. The only thing you can say is that they make it convenient to use with their tool chain.

Re:what a wimpy database

[zardo]

Basically the same as what Sybase developer edition is. I think Oracle is a bit better than Sybase for speed, feature wise it's about the same I'm sure.

Re:what a wimpy database

[zardo]

Since I can remember Sybase has had a free version of their database. I've always recommended it to all of my clients I do contract work for. Postgres is alright, but it doesn't have the quality feel to it that I get from commercial databases, postgres feels hackish to me. MySQL on the other hand has quality CLI, it's fast, but feature wise it sucks arse. So maybe I'll have something better than Sybase now.

Re:what a wimpy database

[davegaramond]

God, I'll take psql command line client than mysql's any day!

Re:what a wimpy database

[Anonymous Coward]

Read the article -

"It can only run servers with one processor, with 4GB of disk space and 1GB of memory"

How useful can it be if its encumbered like this ? MySQL, Postgres, Ingres, and other open source databases don't ship with these limits.

Re:what a wimpy database

[Dan_Bercell]

Yes it is free, I dont know where you got your information from, but you can freely download it or incorporate it in your commercial sofware (read the license agreement). I think it ships with Sql Server 2000 and Office, but its just a means of distribution. You would be suprised how many SMB solutions that use it, now that Oracle provides the same type of database I am sure it will follow the same road.

Re:what a wimpy database

[Dan_Bercell]

I like it, its nice to not have to pay for the full blown Sql Server edition if its not needed. When vendors only give you the option of MSDE or Sql Server for their application, its nice to save the expense Sql Server licensing. I have been seeing some vendors that give the option to use MySql or other OS database engines, but 3-4 years ago this wasnt even an option.

Re:what a wimpy database

[hey!]

Yeah, there's no question that it's a very convenient and enlightened policy on Microsoft's part. I cannot fault Microsoft in the way it treats developers -- at least up to the point they stumble on what looks like a pot of gold.

Re:what a wimpy database

[Anonymous Coward]

As I recall though, the embedded .dll is only available under Windows (no .so equivalent?) I'm not sure why. Maybe they figure that under linux, you'd just run the full server process?

The embedded version is called Classic on Linux :-) There wasn't Classic for Windows until Firebird 1.5, but Classic implementation on Windows doesn't support "local" connections (which means embedded), only TCP/IP connections. That's the reason why separate embedded version for Windows was created. To make it even more confusing (or interesting), this Windows embedded version has in fact Super Server architecture, not Classic, which is the reason why Windows embedded can't share databases with other server instances (Classic can share). Anyway, the outcome is almost the same, you have a full featured server in shared library.

It Could Backfire

[obender]

Unless Oracle puts together a better administration interface than the current bunch of tools people might actually learn to stay away from it.

Re:It Could Backfire

[Anonymous Coward]

I'll never miss what I never had.

Re:It Could Backfire

[mysticwhiskey]

If anyone and his dog could administer an Oracle database, what will happen to the professional Oracle DBA's? THINK OF THE DBA'S!!! ;)

Re:It Could Backfire

[YoungHack]

> Unless Oracle puts together a better administration interface than the current bunch of tools people might actually learn to stay away from it.

Boy that's no kidding. I've used their real database, and there's no way I would voluntarily choose it for any project of my own. Free wouldn't make a bit of difference to me.

Re:It Could Backfire

[ducttapekz]

I found that managing Oracle's databases is pretty easy. There console is nice and everything is easy to find and intuitively placed. Can you elaborate on what you don't like?

Re:It Could Backfire

[cnelzie]

That would definately be thinking of the DBAs.

    Have you ever seen a Database "constructed" by someone who knows nothing about Database design?

    Easier to use tools will make it all the much easier for Oracle DBAs to quickly repair and rebuild or begin anew a DB created from a person with a Microsoft Access 101 class under his/her belt.

Re:It Could Backfire

[rabel]

Yeah, well I am an OCP DBA from version 8 through 10g and I'll be the first one to tell you to *keep your mouth shut* It's a piece of cake to manage these databases and the big DBA secret is that any moron can do this job. Oracle goes out of their way to make it all completely freaking whacked-out complicated to do anything, but it's still just command line nonsense. Spend 15 minutes reading the concepts manual, pirate a copy of T.O.A.D. and you can be a DBA too.

You wanna know what it takes to be an Oracle DBA? The first 2.5 days of any Oracle DBA class is simply learning how to repeat the mantra of the DBA. Ok, everyone say it with me now:

It's not a problem with the database.

I suppose I should turn in my OCP membership card now since I've spilled the beans. Go on, try to look your DBA in the eyes without laughing at him the next time he gives you the DBA mantra.

Re:It Could Backfire

[oni]

Have you ever seen a Database "constructed" by someone who knows nothing about Database design?

That would describe 90% of the databases I've ever seen. Then people are amazed when they realize that there are questions that their data CANNOT answer, not because the information isn't there, but because of they way they've organized it.

I'll give examples if anyone is interested.

Re:It Could Backfire

[kpharmer]

While it's true that keeping a tiny oracle vanilla database running can be very simple - especially if you don't care much about the data, and aren't developing on it (canned app).

It is not simple to admin if:
    - you've got development trying a lot of different ideas - and you're spending all your time researching issues
    - you've got a ton of data
    - you've got business critical data and need to configure for maximum reliability
    - you get into a slightly off-the-path data restoration scenario

And the work of a dba isn't trivial if you are trying to do more than just keep the server running. In that case you should be:
    - handling data quality issues
    - handling data concurrency issues
    - handling multiple apps using the same data elements
    - handling replication issues
    - designing reporting models
    - often implementing the ETL process between operational and reporting models
    - performance-tuning databases that are constantly changing over time, often selecting the new hardware, then migrating them to new hardware.
        Figuring out where to spend your money to get performance you need: more memory? more cpus? more/faster disk?
    - writing queries for developers that don't know how to write simple sql
    - writing queries for advanced developers that aren't sure how to get the best performance out of complex queries
    - creating automatically-maintained summary tables
    - researching fixpacks and upgrades to figure out which you need, and what they'll break; then doing
        emergency un-installs or work-arounds when you're wrong.

If your dba job is simple you're probably doing about 1% of what dbas do at our location

Re:It Could Backfire

[frisket]

Not just the admin; if the installer is as badly-written as the one that installs the Small Scale Oracle that comes with Blackboard 6.1 Basic (to date my only exposure to Oracle) then Oracle Corp has a long way to go.

It looks like it's been written by someone who has heard installation routines described, but never actually seen one.

And if it's not Oracle's fault but Blackboard's (unclear from the software which) then Oracle still need whipping for allowing a partner to do this to their product.

By lagging so badly behind with XML and UTF-8, MySQL risks losing territory to Oracle, who stitched up the corporate datacenters a decade ago. Cost and openness are not the factors here: the important ones are trust and scalability and someone to finger when it goes wrong.

Re:It Could Backfire

[secolactico]

Since we are sorta on the subject, can anyone recommend a book or collection of tips about database organization for a beginner?

I've done sone (very) lightweight database tasks for transient purposes and would like to acquire some "good habits" before I delve any deeper. This would be more for hobbyist purposes than for a production environment. I've used Mysql, Postgresql, MSSQL and DB2.

Regards

Re:It Could Backfire

[oni]

You might try some university-level database books. The advantage there is that they tend to concentrate on the underlying principals rather than any particular database implementation. Since you already know the syntax of the database you want to work with, you should be able to apply whatever you learn in the book to what you're actually doing.

The textbook they used in my "Database 101" class is still available on Amazon. It's ok. I suppose it's probably about average.

Title = Database Design, Application & Administration
Author = Maminno
http://www.amazon.com/exec/obidos/tg/detail/-/0072 880678/103-4666631-1635844?n=507846

If there is a university close to where you live, you might also try going to their bookstore and looking at the used books they have.

Just stay away from those, "Learn Microsoft SQL Server 2005 in 21 days" books. That's the trap many people fall into. You don't want to learn a particular implementation. You want to learn normalization. I promise you, it will make your life easier in the long run. Normalization is to database design what goto-less programming is to programming. When you start programming, you do it on your own and you learn bad habbits, like using gotos all over the place. Then one day you sit down with a book, not a "C" book or a "Java" book, but a programming book, and that's when you really learn. Same here. You want to learn design and normalization.

good luck. I think it's good that you're motivated to do this.

Re:It Could Backfire

[aralin]

Actually, with Enterprise Manager and all the different Advisor packages, everyone and their dog and their cat and even their grandmother can administer an Oracle database. Its scary to even imagine, but what can happen, will happen.

Re:It Could Backfire

[EnronHaliburton2004]

I can't even install the damn thing. The installer (Oracle 9i) crashes 75% of the time-- if I forget to set the "fileHandlerFooBar" property to 95,000 files (or something), then the Oracle installer should COMPLAIN and TELL ME, but instead it encounters an unexpected exception and crashes. Beautiful.

The fact that the 9i installer uses Java 1.1 and I need to tweak a dozen low-level OS settings just to install Oracle does not inspire confidence. This makes sense for a high-end production database, but it doesn't always make sense for a development machine.

I shouldn't need a DBA just to install the software on a test machine.

Re:It Could Backfire

[orderb13]

ANY book that talks about a database should have a chapter about Normalization in it. Or you can just google "database normalization".

Remember boys and girls, third normal form is, mostly, your friend.

Re:It Could Backfire

[nettdata]

The new installers for 10g actually DO tell you if/when some kernel param hasn't been set, etc.

You're right, though... Oracle typically has a VERY shitty track record with their installers. I remember taking 3 days to re-write the Apps 11i installer so it'd work (even though the installation itself took over 12 hours). The local Oracle tech's were begging for my re-write, as THEY couldn't get it installed and operational.

I'm still pissed over the fact that they got rid of their command line installer, among other installer issues. (Like, why oh why do you need a graphical environment to perform a silent install? It's like the marketing department designed the installers)

That being said, I'm still a big fan of using Oracle for medium to large projects, and find that most of the "developers" that bitch about or rag on Oracle DB's are either working on very small/simple apps where mySQL is addmittedly the tool for the job, or they just haven't used it before, or they just don't understand it.

Sure, I tend to use Oracle for everything, but that's probably due to the fact that I run an Oracle Dev shop, and we've got crap-loads of the things installed and running all over the place, and have all the build and monitoring tools and developer environments to work with them faster than anything else. I'm also probably pretty jaded to any/all of Oracle's foibles, as I'm used to dealing with them so much that I don't notice them any more. ;)

Re:It Could Backfire

[iabervon]

I think it's gotten somewhat better in the past few years, at least from what my DBA friends were telling me. Anyway, there are a number of database administration programs that work on an Oracle database in addition to other databases. IIRC, hsqldb's front end will work with any database that supports JDBC, and the professional DBA I know was using Toad to administer Oracle last time I checked. That's the nice thing about standard APIs: you can choose different things for each side without too much hassle.

Re:It Could Backfire

[EnronHaliburton2004]

I'm still pissed over the fact that they got rid of their command line installer, among other installer issues. (Like, why oh why do you need a graphical environment to perform a silent install? It's like the marketing department designed the installers)

Generally my production machines don't have any graphical capabilities-- No X, no monitor, videocard or keyboard. Ethernet and a Serial Console provide all the connectivity we need.

If a GUI is required for installation, can I install Oracle on these machines? Does Oracle allow you to do a remote install?

Re:It Could Backfire

[benjamin264]

_An Introduction to Database Systems_ by C.J. Date.

http://www.amazon.com/exec/obidos/tg/detail/-/0321 197844/qid=1130779484/sr=8-1/ref=pd_bbs_1/102-1718 741-7424931?v=glance&s=books&n=507846

I think that the greatest thing Date ever said was, 'The principles of database design are nothing more than formalized common sense.'

Re:It Could Backfire

[nettdata]

Their installer requires the graphic libraries in order to launch, even if no graphical components are required. So, even if you perform what they call a "silent install", where you just launch the installer and pass it a parameter file, it still needs the graphical libraries in order to launch.

This has been a HUGE issue for some of the Government installations I've done where X is not allowed, and they've not even installed the basic libraries. It's also been a VERY stupid scenario, in my opinion.

In these cases, we've used a virtual frame buffer and/or VNC to do the installs. All of my DB servers are headless, and I just tunnel VNC over SSH for the installs.

Personally, I prefer the VNC approach, as it allows me to connect in, do what I need, then not have to worry about losing a client connection... the session is resident on the server.

Re:It Could Backfire

[jsantos]

Actually, I'm interested, to tell the truth I haven't had the misfortune of dealing with badly designed databases and I would like to know what it takes for people to make them unable to provide answers while the information is there.

Thanks in advance,
Jorge

Re:It Could Backfire

[frostman]

I'm a database designer and have had similar experiences... and my skin crawls just reading your post.

OK, so how about some examples? At least one, in the spirit of the Daily WTF :-)

Re:It Could Backfire

[oni]

What I've seen a lot of are databases designed by programmers - good programmers too, these guys weren't slouches. They didn't have any formal training in databases or maybe they hadn't paid attention. Programmers tend to approach database design by thinking, "what data structures do I need to get out of this?" Someone more experienced in database design approaches the problem by thinking, "what relationships are here and how do I model them?"

Here's a real-world example. A web-based application that I was hired to extend and maintain included a system for users to exchange the lesson plans they created on the site with other users. In doing this, the users built up something like a buddy list. "These are the people that I often share with." Or you can think of it as being like an address book.

With me so far? OK, the programmer (who was very sharp - probably better at this than I am) approached the database thinking, "what data do I need from this thing," and decided that what he needed was a comma delimited list of userIDs. So he physically stored the buddy list in the database in like a char(500) as a comma delimited list.

That was actually great for what he was doing. He was just showing a user their buddy list. Unfortunately, that isn't normalized. So, there is a question you can ask which can't be answered by the data. That question is, "how many people have userID 50 as one of their buddies?"

See, the correct way of doing this is to have a many-to-many relationship which you implement with a table containing just two columns, userID and BuddyID. So if I'm user 12 and users 13,14, and 15 are my buddies, I have three rows in that table:

UserID BuddyID
12 13
12 14
12 15

Now if I need to ask, "how many people have userID 50 as one of their buddies" I can do select count(*) from X where BuddyID = 50

BTW, I actually fixed this one not by normalizing but with a hack. I appended 0 to the front and back of the buddy list, then I could do select count(*) from user where buddylist like '%,50,%' But hacks aren't how I make my living. I'd prefer to do things the right way. What my boss wanted was, every time you look at this page it shows you who has you set as a buddy. Kind of like what Slashdot does with the "fans" page. If it was normalized, that would be a scan of an indexed column. It would be lighting fast. so fast the page would practically load before you even clicked the link. But doing a "like" on a big char field is slow.

There are I'm sure still other questions that the un-normalized database cannot answer. Also there are problems with deleting users and, the big one, overflowing that char(500). "How many buddies can a person have?" I was asked. "It depends" I said. If all my buddies are low number IDs, I can have a lot. If my buddies are high number IDs, I can have fewer. It's all just huge mess! Of course, it worked according to the original specification though.

I have a lot of respect for the guy who wrote it. And I'm not tooting my own horn either. I've have other people look at databases I've designed and just torn them to pieces. This is just one example where I just happen to know that he did it the wrong way.

Re:It Could Backfire

[kpharmer]

> Actually, I'm interested, to tell the truth I haven't had the misfortune of dealing with badly designed databases and I would like to know what it takes for people to make them unable to
> provide answers while the information is there.

here's a list of major screw-ups:
1. objects are stored as binaries instead of separate columns
2. data is organized in massive and simple key-value pair columns rather than as separate and dedicated columsn
3. data is kept as strings that must be parsed with application code instead of with sql

lesser screw-ups would include:
4. overly normalized model
5. missing data
6. lack of versioning or history
7. data versioning inconsistently applied
8. data quality inconsistent due to all validation being performed at application layer
9. etc

Many of these issues don't prevent users from getting their data back out the application that wrote it. The problem is that when they have a simple question like, well, how many transactions per day have we been getting for the past year, and is the rate increasing? That there's no way to answer that question through SQL, but more expensive application logic has to be built. Then you next discover that the application that may have performed fine for simple transactions *completely dies* when you try to have it scan a year's worth of data.

The root cause of all this is that many people are still building applications like it was 1985 - and just being able to CRUD a few objects is good enough. Well, it isn't. People generally expect to be able to get insights about a business process in addition to simply creating operational transactions within it. And if you don't know that up front, and then design the database in accordance with the latest fads in J2EE, etc then you *will* suffer once that application starts getting used.

Re:It Could Backfire

[rabel]

Oh give me a break. You're missing the point. You and the little "I am *too* important" nimrod DBA who rated my message Troll.

All the things you mention are normal everyday DBA issues, not-specific to Oracle. Well, other than the endless stream of fixpacks and upgrades required for every Oracle installation. I was speaking about how managing Oracle databases isn't all that hard, but that Oracle makes everything much more complicated than necessary. The secret is, there ain't no secret.

Sorry dude, you certainly don't need to be a rocket scientist to manage an Oracle installation. The issues you're referring to are mostly problems for the development DBA, and that's where all the smarts are necessary. But hey, if you're convinced your DBA's hung the moon, more power too them. They've got you fooled. In fact, why don't you download a free development license for Oracle 10g Enterprise Edition and change careers into the DBA role? You'll most likely get a big fat raise and then you'll realize just how easy it really is.

Re:It Could Backfire

[kpharmer]

> All the things you mention are normal everyday DBA issues, not-specific to Oracle.
right

> The issues you're referring to are mostly problems for the development DBA, and that's where all the smarts are necessary.

This might be part of the confusion: don't separate production vs development dba roles. You're right - that a production dba role can be very simple. So why even bother with it? Unless you've got complex security requirements for it, you are far better off without that role.

So then, back to the dba role. Yep, you need to be able to write complex sql, design a data model, tune the database, select the hardware, and plan and administer a data recovery procedure. And yes, it is a hard job.

Re:It Could Backfire

[einhverfr]

FWIW, this is probably the most authoritative book on the subject.

Re:It Could Backfire

[rabel]

a production dba role can be very simple. So why even bother with it? Unless you've got complex security requirements for it, you are far better off without that role.

Those production DBA's that do not have their head in the sand after the release of 10g know the score. Production DBAs are a dying breed. Oracle databases will eventually be self-managing and self-tuning in the very near future.

But as you say: complex SQL, data model design, ETL, Data Quality... you betcha, those roles won't go away until we've got either AI, or we rid the planet of users. Anyway, I was only being a smart-aleck in my original response. What ticks me off is the "know-it-all" Oracle production DBA that tries to tell everyone what he/she needs, rather than the other way around. Since everyone thinks the DBA actually practices computer-voodoo in order to keep the mysterious and all-powerful Oracle database up and running, they can sometimes get away with it. I'm here to tell you that is just ain't so. It's not hard, it's easy to learn, everything you need to know is in the docs with step-by-step tutorials online and the tools make it as easy as point-and-click.

Re:It Could Backfire

[zeke2.0]

Unfortunately the new msde coming out will have limited admin tools too. Sorta sucks IMHO. Btw there is an excellent and relatively affordable admin/developers tool called PL/SQL developer at this site http://www.allroundautomations.com/ they've been doing this stuff for many years, and it's a viable alternative to TOAD for all but the uber/master Oracle DBA types. Under 200 dollars US. Has a free fully fucntial trial download.

Re:It Could Backfire

[Al Dimond]

Wow. This hits close to home for me... I don't have any formal database experience, and a small amount of practical experience, and this summer I had to work on a database app. In programming terms there were two data structures, that had to be linked in a many-to-many fashion. My first thought was that one should have a linked list of pointers to the other (well, it would be a linked list of IDs, not pointers). Can't do that. An array of pointers? Nope. Hmmm, I thought long and hard. Well, I thought, for this particular application we figured we'd probably only need about 100 records of the second type, so I said, AHA, I've got it! I'll use a bit vector! One bit for each record of the second type! So I put 4 integer fields in the first table and started to code up the application. (mods that have never used databases are probably thinking -1 "redundant" or +1 "interesting". mods that have used databases are probably thinking +1 "funny" or -1 "if ya like C so much why don't ya marry it")

But it just felt like the Wrong Thing, so I asked my boss and he told me to google for many-to-many tables. And now I kind of get the idea of relational databases. Sometimes I even see ways to use that kind of organization in C-style programs. I don't use it a lot, but sometimes it gives me another way to look at problems, which is helpful.

Re:It Could Backfire

[fbg111]

I second the recommendation for C.J.Date's book. After you finish reading that one, here are some more, from Date, Darwin, Pascal, and other standard bearers of the relational model.

Re:It Could Backfire

[Anonymous Coward]

EXACTLY - It's a thought process. A way of thinking out a solution. My job title is Oracle DBA, one of two. I spend most of my time fixing code from the previous "DBA". A lot of the problem's are caused by the schema design (if you can call it that). Some schema's do not have ANY FKs! Then there is the processing procedures. They all use looping to process the data 1 record at a time. SLOOOOWW. For example, uploading 200 records from 10 or 11 tables into about 15 tables takes between 6-8 hours. With error logic that is flawed no less. I am just starting the re-write (from scratch), but I figure to get it down to 4-5 seconds per 100 records (There is a back log of about 15,000 records).

FYI - the previous "DBA" was a Peoplesoft Administrator. Smart guy, no concept of programming (expecially DB programming). But I do feel that this can be learned if one desires to. Read the CONCEPT guide. Also the SQL guide. Then onto PL/SQL. Those three guides will take a beginner at least a couple weeks to read and practice. They will recoup that time on thier first project alone!

Re:It Could Backfire

[kpharmer]

> Those production DBA's that do not have their head in the sand after the release of 10g know the score.
> Production DBAs are a dying breed. Oracle databases will eventually be self-managing and self-tuning in
> the very near future.

Yep, they're trying to get there. It's a long haul from where they started, but the production dba role will be 95% automated in the near future.

It's kind of like IT operations staff. In the old days we always had operators to manage the daily manual tasks of IT. Now this is viewed as a crutch - things should be *so* automated that operators aren't necessary any more.

> What ticks me off is the "know-it-all" Oracle production DBA that tries to tell everyone what he/she needs, rather than the other way around.

I think this is a common symptom of administrators who are too far away from the business. Whether system, network, database, or applications - they guys often start to think of themselves at the center of the universe - instead of serving some business need.

Re:It Could Backfire

[zardo]

Yeah, not without a crapload of scripting. I have seen too many databases where the architect originally thought that the fewer tables the better, even if it meant widespread replication. I once worked with a database that was essentially one table. The users login and password was in the main data table, replicated dozens of times in some instances. I ended up turning that one hulking table into 10 or 15 different tables. That is the classic mistake, fewer tables is better. I think it's because they don't know how joins work or something.

Re:It Could Backfire

[Anonymous Coward]

With me so far? OK, the programmer (who was very sharp - probably better at this than I am) approached the database thinking, "what data do I need from this thing," and decided that what he needed was a comma delimited list of userIDs. So he physically stored the buddy list in the database in like a char(500) as a comma delimited list.

If the programmer really did this, then he was far from being "very sharp." It's just plain stupid, and any programmer with half a brain knows that.

Re:It Could Backfire

[Ashish Kulkarni]

I'm curious, could you show such examples?

Re:It Could Backfire

[oni]

I'm curious, could you show such examples?

http://slashdot.org/comments.pl?sid=166851&cid=139 17525

Re:It Could Backfire

[Anonymous Coward]

Anything by CJ Date, but start out with "Database in Depth" on O'Reilly. A very slim book with a lot of good information on the relational model (the ONLY complete model for data management).

I can't stress enough how well Date understands and conveys the information. You won't find clear thinking on databases anywhere on the internet (except maybe at dbdebunk.com), and even most books muddle the basic concepts. It's really mind-blowing a) how simple the relational model is b) how few people have even tried to understand what it is and what it means (hint: it's not SQL).

Be one of the few that *understands*.

Nice TRY

[LogicallyGenius]

But, I will go for Open source and free of any license; so which should I go for ?

Re:Nice TRY

[LLuthor]

PostgreSQL is your best bet in terms of Oracle like features with an Open Source (BSD) license.

Some very nice Oracle features are also being developed in Bizgres.

SQL For Fun?

[Goo.cc]

I've been meaning to ask this for a while on Slashdot, but how many people here use a SQL database as part of a hobby or for fun?

I currently have PostgeSQL running on my Tiger box. I initially installed it just to experiment with SQL and database normalization, but now I keep my comic book inventory on it. (I know that this is like swatting a fly with a nuclear weapon but I enjoy using PostgreSQL and it is FREE software.)

As for Oracle's announcement, I think that it can be a good thing, provided you are willing to live with their restrictions and only need support for Linux (x86?) and Windows.

Re:SQL For Fun?

[FinestLittleSpace]

I think any coder with a bit of common sense uses SQL for personal use. It's so diverse and so easy to use. I used to run a large music site off the back of SQL, but have used SQL for all sorts of easy storage. Put on a local server with phpmyadmin it's really useful.

Re:SQL For Fun?

[Dystopian Rebel]

I initially installed it just to experiment with SQL and database normalization, but now I keep my comic book inventory on it.

Welcome to /. O Kindred Spirit! Your journey has been long, but now you have found us and you may rest.

I know that this is like swatting a fly with a nuclear weapon

That's a game scheduled for release at Christmas.

Re:SQL For Fun?

[will_die]

MS-SQL server and/or Oracle are one of the first applications I install on a new computer. However I am a DBA and database developer.
As for what this will do for Oracle probably not that much. You could already could get a free licensed copies for development purposes, this allows distribution which is nice. This will be nice for the few times when you need to include a server based database with your application, but unless it is major application with multiple users give me a file based system and let me ignore the overhead.

Re:SQL For Fun?

[Eil]

I know that this is like swatting a fly with a nuclear weapon...

Not at all. As long as PostgreSQL enables you faster and easier access to your inventory, then that's reason enough to be using it for that purpose. A lot of people are under the assumption that a full-fledged database is only appropriate for large or complex data sets. This is wrong. I use PostgreSQL on my home server to hold contact information for a "friends and family" addressbook as well as a few other teeny-tiny datasets because it was simple to set up, is trivial to back up and maintain, and can be easily exported in practically any way. You're not alone in using a database for what most consider relatively minor tasks. :)

Re:SQL For Fun?

[everflow]

I've been using databases for personal use for years. Started with Windows/VB6/MS-Access (yeah I know but everybody has to start somewhere) and ended up with Linux/Java/PostgreSQL. It began with the little thought that a database would be nice to archive and connect all my little lists (personal address book, jobs, projects, institutes, journeys, books i read, ...) and it grew and grew.

I'ver learned a few interesting things:

- Organizing your data in a database makes lists go 3D. I now can see connections (with the right db-architecture and SQL-knowledge) that I couldn't have seen with looking at lists.

- Almost all of the tables are connected to the list of people I know (journeys I made with friends, books someone advised or borrowed me, money I owe or lend, projects/jobs I did with other peoples ...). I suspected something like this in the beginning but it was fascinating to see how centered the contact-table ended up.

- Although it took quite a time it was worth it because I use it almost at a daily bases to look things up.

- Instead of having less lists outside the db I have many more. Their number exploded but at the same time I learned so much about structuring my infos, that they are still manageable and useful (with spreadsheets) ... it's like an external harddrive for my brain - faster and much more secure against loss of information.

I have to admit I can't imagine living and working without my omniscient database anymore ... maybe that's one of the reasons why I feel so at home at /.

Re:SQL For Fun?

[Daytona955i]

I started playing with mysql, doing little things here and there but nothing serious. I used that knowledge of SQL to get a job doing development work on a set of programs that use an Oracle DB as it's backend. Right now I am working on my own web-based (and once that's finished I'll write a standalone program) to provide family and friends access to my recipes.

So I use SQL for both work and for personal use. Though I will say, even though I would meet the requirements for the free Oracle DB, I know what an Oracle DB is like and I would much rather stick with mysql or PostgreSQL thank you very much.

Re:SQL For Fun?

[brlewis]

I use PostgreSQL for my Toastmasters club hobby, letting people sign up for meeting roles online.

Only in a really twisted sense is keeping your comic book inventory on it overkill. You could just as easily say it's overkill to keep your physical comic books in your house if a locker would do. But if you have a house, keeping your comic books there is the most convenient thing. Same with a real database. Once you have it there are advantages to keeping everything there.

Re:SQL For Fun?

[plopez]

I have been playing around the Postgresql, trying to integrate it with GRASS http://grass.itc.it/ (and learn Grass 'on the fly' while I am at it) and warehousing water quality data. Sounds like a cult may be forming :)

Re:SQL For Fun?

[droleary]

I currently have PostgeSQL running on my Tiger box. I initially installed it just to experiment with SQL and database normalization, but now I keep my comic book inventory on it. (I know that this is like swatting a fly with a nuclear weapon but I enjoy using PostgreSQL and it is FREE software.)

SQLite goes one better and is public domain software. It comes with Tiger, too, which makes it a great tool if you don't need an actual server for your SQL play.

Re:SQL For Fun?

[Max Threshold]

My first foray into DB design was a PHP/MySQL site to support a meme on LiveJournal. I got 500,000 hits on the first day, and took down the entire shared server. I guess my $4.95 a month hosting wasn't designed for that. :o)

Now I'm making my living designing a Postgres DB for a major telco.

hopefully..

[icecow]

'hopefully' is code for 'by then they somewhat locked in'

Restrictions?

[mysticwhiskey]

I guess it all depends on the restrictions applied to the "free" version. Is it crippled in any way? Does distribution of the free version require certain conditions (ie development criteria, use of application, etc)?

The article states hardware restrictions of "one processor, with 4GB of disk memory and 1GB of memory", so this may automatically disqualify applications of a certain scale.

Move along, move along ...

[Daniel Dvorkin]

... this is crippleware. It's no threat to MySQL, PostgreSQL, or any other open source DBMS, because the developers of those databases are working to put as many features as possible into their free products, while Oracle is deliberately taking features out. This will probably be a good resource for people who want to learn Oracle on their own time, or organizations already using Oracle that want to test a new rollout without having to pay additional fees via Oracle's baroque pricing scheme, but that's about it.

Re:Move along, move along ...

[Anonymous Coward]

Anyone wanting to learn Oracle can and have been able to download the latest versions of all their software for free anyway!

Re:Move along, move along ...

[popeyethesailor]

Well, Oracle has always been freely available for non-production use.. They even mail out their entire range of software(DB, App server,dev tools etc) free of cost.

I once received 10g for linux, and the box had every latest release of Oracle software for Linux. They're quite developer-friendly; just as MS is. For production use however..

Re:Move along, move along ...

[Daniel Dvorkin]

Sounds to me like the difference is that you're allowed to use this new edition for (limited) production use as well. Now, I'm sure there are a ton of small shops currently using the free, "non-production" edition for production apps, but of course they're not really supposed to; this gives them a legal route. But I still think the deliberately crippled nature of the product makes it unattractive relative to the open source contenders, in terms that even PHB's will understand: "Boss, if we go with 'free' Oracle, we're going to run into that disk space limit pretty fast, and then we'll have to pay $$$."

Re:Move along, move along ...

[pci]

Oracle is only free for non-production use if you are in ISV.
If you do in house programming, you are still suppose to buy licenses for Development.

Re:Move along, move along ...

[KiloByte]

The problem with Oracle is, it doesn't scale at all. It is meant to do grid computing, but can't really do anything smaller.

How often do you need to use a cluster for your data? If you are a major organization, then you will, but the majority of installations are pretty small. Firewall/website logs. Customer data. And so on.

I have once developed a workshift-tracking application for a company with around 200 employees. A couple of years later, the total data takes 17MB. Why would you use Oracle if MySQL works faster and takes 1% of the resources? A minimal installation of Oracle 10g takes ~800MB of memory, and will take over ten hours to install on a machine with 512MB ram, on the other hand, on my firewall (486, 32MB ram) MySQL can handle Apache logs (only about 200k hits, though) taking a split second for any reasonable query.

Oracle works better for clusters.
MySQL works better for a single machine.

MySQL is a lot faster. Oracle takes distributed processing a lot better.
But uhm, where does a crippled version fit in the picture?

Re:Move along, move along ...

[LLuthor]

This is probably a good way to learn Oracle's nuances and help you on your way to become a DBA for Oracles bigger stuff.

This is not in response to OSS (since they target two distinct markets: MySQL low-end, Oracle/DB2 high-end). Rather, this is likely in response to the MSDE (and the SQL Server 2005 version thereof), and the free DB2 products.

This version is licensed for commercial use (unlike all the other free stuff oracle gives away), so they are probably hoping people will develop for this and eventually migrate up to the higher-end products offered.

Re:Move along, move along ...

[Varun Soundararajan]

so.. this is free as in beer! ;)
=======
This space is now not blank :)

Re:Move along, move along ...

[rjstanford]

I have once developed a workshift-tracking application for a company with around 200 employees. A couple of years later, the total data takes 17MB. Why would you use Oracle if MySQL works faster and takes 1% of the resources? A minimal installation of Oracle 10g takes ~800MB of memory, and will take over ten hours to install on a machine with 512MB ram, on the other hand, on my firewall (486, 32MB ram) MySQL can handle Apache logs (only about 200k hits, though) taking a split second for any reasonable query.

Oh, come on. We all run Oracle on our development machines, and on many of our servers (Windows and Linux both). Configured to provide "adequate" performance for OLTP on ~10G of data, it takes about 256M of RAM (some swapped out, of course). I could probably get that lower, but it hasn't been a priority (a ton of settings are left at their defaults). Installation takes about an hour on each platform - too long, I agree, but not the all day affair you make it out to be.

I don't personally like using it because of the lack of standard tools, but hey, I'm an old school Informix guy myself. Still, its readily available, it works well, and once you master a few ideosyncracies its not that hard to deal with. So lay of the FUD a little, would ya, or at least come back with some more belivable numbers. I'm not disagreeing with you that MySQL could work fine, note, just that there's no need to exaggerate its benefits like that.

Re:Move along, move along ...

[kpharmer]

> But I still think the deliberately crippled nature of the product makes it unattractive relative to the
> open source contenders, in terms that even PHB's will understand: "Boss, if we go with 'free' Oracle,
> we're going to run into that disk space limit pretty fast, and then we'll have to pay $$$."

Actually, it looks to me like it will be extremely attractive to quite a few people:
    - can easily handle small or departmental databases
    - very cheap (free)
    - completely consistent
    - same codebase as oracle 10g
    - should enable easy portability for oracle-enabled applications
    - should be able to easily leverage existing oracle & developer skills

While it is true that this version has extreme limitations (1 CPU, 1 gbyte memory, 4 gbytes data), and if this is your *only* database server it may not be a good choice. But on a 3ghz server with 15k disk and good disk cache it could be extremely fast for small projects. Meanwhile, if you've got or need a dozen databases in-house, ranging from small to large, this solution could keep the small database costs down - while also eliminating the labor costs associated with managing a variety of products. Further, if the occasional project needs to go up from 1 CPU to 2 - the licensing costs are fairly cheap. Going from 1 CPU to 64 is a completely different situation (could cost a ton), but then again think of the massive labor savings.

I'm currently involved in leveraging this exact same strategy with db2 (their cheap express product came out about two years ago): we try to migrate all databases to db2 to use consistency to keep labor costs down. Then we choose the database license that'll be the cheapest to purchase and manage. Conversions from mysql are sometimes a pain in the butt due to weirdness of that product - then again, its data quality problems usually make it worthwhile anyway. Postgresql is more complex - the quality of this product is fine, the only benefit to conversion is administrative consistency.

Anyhow, is this product a mysql-killer? No, but could definitely cut into its market:
    - ISVs with small databases could use this instead of mysql with no problems
    - it is freer than mysql
    - ISVs that need more power in their product beyond this can easily pick a more powerful version of Oracle that can completely outpace mysql anyway
        (assuming you use parallel capabilities). This will cut down on hardware and development costs.
    - mysql's future is now completely up in the air since Oracle has bought out the best part of mysql (innodb)

Of course, postgresql and firebird are still out there. These products are healthy and steadily improving, just lack the market share to compete well with oracle right now.

Re:Move along, move along ...

[electroniceric]

Oracle also has a number of certifications that make it very attractive in certain industries like defense, healthcare, etc. I work in the medical field, and a free 4GB-max "validated" Oracle database is a huge boon to a wide variety of medical ISV's. Given that's it free to redistribute, the OSS/proprietary thing kind of takes a backseat for a lot of these ISVs.

By the same token, I also don't think it's going to drive sales in the way they think it will. Databases are slowly but surely going commodity, at least at the lower end of the market, and this merely reinforces that trend. And along with that, there's an increasingly robust set of tools to obviate the differences between these database for most uses that don't demand extreme peformance, from Hibernate and ORM packages to ADODB and other database-independence layers in PHP to .NET's layered data architecture.

As a Postgres user, I'm hopeful that Sun's proclaimed interest in Postgres will result in this kind of "validation". However, given Sun's reputedly somewhat lackadaisical commitment to staffing OOo, I'm not holding my breath. With Postgres' extensibility and extremely high-caliber core developer base, I think a strong commitment to validation by Sun could make it a real contender in the medium enterprise space. Validate it, clean up a few features (notably auto-vacuum and passable auto-tuning, maybe some multi-master replication), throw in a simple deployment for ORM or database indirection, and you've effectively moved that commoditization up one layer from the small website developer level.

In the long run, I don't see how this gets Oracle out of the need to transition its core revenue off of its database licenses.

Re:Move along, move along ...

[briansmith]

Sorry, I have to call bullshit on this. The reason Oracle takes 800MB of memory is because you TOLD IT TO take 800MB of memory. It will work with a lot less. I remember that I once saw a departmental database that had Oracle limited to like 150MB of memory and it ran suprising well considering there were 30+ people using it at a time.

You say that Oracle doesn't scale at all. Normally when we say "doesn't scale" we mean "doesn't scale up," but you seem to mean "doesn't scale down," since saying Oracle doesn't scale up is laughabled. But, for really small databases there is "Oracle Lite" which runs even on PDA's.

You say that Oracle takes up a whopping 17MB of disk space for your data. Is it 1994? Why is this even an issue? Honestly, I am suprised that it takes so little, considering you had to set up temp tablespaces, logs files, etc. When I was DBA'ing I think each of my log files was 50MB.

Re:Move along, move along ...

[Anonymous Coward]

Well, if you bother to actually try it out, you'll find that it is tons easier to install and comes with a nice web-based administrative interface. I'm recommending that my developers consider this for an at-home development system since it is so easy to use that I have no doubt they'll be able to handle it without bothering me.

Free has very little to do with why I think this is a good idea.

Re:Move along, move along ...

[Eivind Eklund]

MySQL does not provide the feature of "When I insert data into the database the same data comes back" (due to various "helping" features of the frontend.)

Due to a combination of the number of bugs and misdesign, MySQL does not (to any reasonable degree) provide the feature "When I do an update to the database and do not get an error, that update is guaranteed to be there". It happens regularly that it isn't there.

These are really core features for the kind of developer I want to be. I'm really, really sad to be working with MySQL day to day - it makes me feel dirty.

Eivind.

Re:Move along, move along ...

[javaxman]

Due to a combination of the number of bugs and misdesign, MySQL does not (to any reasonable degree) provide the feature "When I do an update to the database and do not get an error, that update is guaranteed to be there". It happens regularly that it isn't there.

One of the many, many reasons why we chose PostgreSQL.

If I thought you had any choice in the matter, I'd ask why you aren't using it. Short of the "it uses less resources" argument, I've always been confused as to why you'd choose MySQL when PostgreSQL is available... not that I'm not open to using MySQL... I just never found the arguments for it too compeling, and too often they're like "it works for me, what's your problem, loser!", which doesn't really help MySQL in a corporate or even small business setting.

No, this shouldn't be a PostgreSQL/MySQL debate. We should be talking about this Oracle thing... which is nice, but just reflects matching what the competition already offered, right?

Re:Move along, move along ...

[KiloByte]

You say that Oracle takes up a whopping 17MB of disk space for your data. Is it 1994? Why is this even an issue? Honestly, I am suprised that it takes so little, considering you had to set up temp tablespaces, logs files, etc. When I was DBA'ing I think each of my log files was 50MB.

Eh? I meant 17MB of _data_, not 17MB of disk space taken. The server which handles the database is (IIRC) a P4 with a RAID1 160GB disk; just a random ok machine at the time. It serves around ~15 clients for two applications.

My point was, it's a typical company and it doesn't produce excessive amount of data. With 200 employees, several records per employee who worked in production on a given day and a couple of years, the whole database doesn't require a grid.

Also, I am sure you can tune an Oracle server down. My whole experiences with Oracle are a semester at university several years ago and setting up an installation for evaluation purposes once. I'm not an Oracle DBA, I just notice that it's too elephantine for your everyday website. If I were setting up an ATM processing datawarehouse, I would know to look elsewhere than MySQL.

Re:Move along, move along ...

[nettdata]

Seriously?

Sounds like you don't know what you're doing. I routinely run MULTIPLE instances of Oracle on my 2GB laptop, each instance taking up less than 120MB of memory.

The biggest performance issue I have with them is the disk IO.

FYI, I run an Oracle shop. We've worked on over 300 different Oracle installations in the last 5 years, on everything from US/Canadian Government sites, on-line banks, to gaming sites (poker, etc) and custom apps. In all that time, we've only worked on 5 clusters, and 2 of them were specifically used for grid type scenarios. (The others were implemented for failover).

If you're expecting to be able to "understand" Oracle and get it installed and running and build something on it in a few hours, then you're looking at the wrong database. If anything, it might be said that Oracle gives the developer/dba TOO much control, and therefore makes it easier for the end-user to royally fsck things up.

I've spent over 15 years working with Oracle DB's (and others) and it is a CONSTANT learning curve. But then again, it's my trade, and not a hobby.

If you want to build your hobby-ish "web site stats" db, then you're probably better off using sqlServer or something more simple.

If you want to do some heavy lifting, then invest some time in learning Oracle, and it'll pay off big time.

Re:Move along, move along ...

[KiloByte]

I routinely run MULTIPLE instances of Oracle on my 2GB laptop, each instance taking up less than 120MB of memory.
If you know more about Oracle than me, sure, you are able to tune it down. To 120 freaking megabytes. This means, a program that uses an Oracle database will waste 120MB of memory just to have its data accessible in SQL form. If you run several different things which require a database, this adds up to something completely unacceptable for a shared server.

What most people need, is a SQL database as a tool, not as the main component. The database is supposed to store data in an efficient form and provide you with an ability to quickly query the data, nothing more.

If you're expecting to be able to "understand" Oracle and get it installed and running and build something on it in a few hours, then you're looking at the wrong database. If anything, it might be said that Oracle gives the developer/dba TOO much control, and therefore makes it easier for the end-user to royally fsck things up.
Good point. I learned SQL (Oracle!) in an university course, then never touched databases in any form for several years. When I finally felt the need, I could learn MySQL and get the job done in a jiffy -- on the other hand, setting up an efficient Oracle server does indeed require a skilled DBA. It's like requiring an engineer to drive a car -- if you want to be able to get from point A to point B frequently, you use a tool that doesn't force you to devote all your efforts into driving itself. On the other hand, if you are a Formula 1 driver, you need a lot of skill. As an another example, a rocket launcher is more powerful than a handgun, but the latter is what most people use, a handgun is a lot cheaper, less bulky and can be easily carried around.

If you want to build your hobby-ish "web site stats" db, then you're probably better off using sqlServer or something more simple.
Or, a system that holds personnel training and machine maintenance data for a mid-sized company. Or, a webapp for an user car dealer. Or, players data for a game with >20k users. Or, billing data for a city-wide ISP. In other words, for tasks where you don't need a full-time DBA just to hold the data.

You will be shocked, but one of the biggest car manufacturers use .dbf files (in dBase3+ format, not dBase4 or never) to hold their data. Not even a relative database, just plain, aeons-old database engine; it does its job perfectly and doesn't require all of the software to be ported. A database is supposed to do its job while keeping its total cost (license+DBA labor+hardware) minimal.

Re:Move along, move along ...

[nettdata]

At the end of the day, it comes down to the requirements of the app and the environment it'll run in, and picking the right tool for the job.

Oracle provides excellent performance, data integrity, security, backup/recovery, clustering, and other features for those people that need them.

If you don't need any of that, or are unwilling to pay for the support environment it requires, then so be it.

I guess my main point was (to put it into your analogy) if you're not an F1 driver, don't bitch because you don't have the skills to drive an F1 car.

Re:Move along, move along ...

[Anonymous Coward]

> You will be shocked, but one of the biggest car manufacturers use .dbf files (in dBase3+ format, not
> dBase4 or never) to hold their data. Not even a relative database, just plain, aeons-old database engine;
> it does its job perfectly and doesn't require all of the software to be ported. A database is supposed to
> do its job while keeping its total cost (license+DBA labor+hardware) minimal.

Having worked at quite a few huge companies (including two car manufacturers) I can tell you that this comment, alone, is worthless:
      - large companies typically have a little of everything due to differences in what various departments order or build
      - they also have quite a bit of legacy systems that can take 20+ years to migrate to newer code
      - they are also heavy users of db2, oracle, and sql server
      - but along with those modern databases, I know they still use vsam, ims/db, adabas, total, and various other misc databases or file managers.

So, are they still using dbase? of course they are. So? It doesn't mean it's a good idea, or that they are happy about it.

Re:Move along, move along ...

[kpharmer]

> Oracle works better for clusters.
> MySQL works better for a single machine.

sorry, no

> MySQL is a lot faster. Oracle takes distributed processing a lot better.

again, no
    - oracle has parallel capabilities that on a single box (2-4 way) allow it to completely kill mysql in performance comparisons
    - oracle has an optimizer that can handle joins of more than 5 tables without collapsing
    - oracle has the ability to rewrite your query to hit an automatically-maintained summary table

The only validity to your claim is that mysql might be a better choice if you're running 1995-era hardware, or are buying new hardware but want less than 128 mbytes of memory (for some unfathomable reason). In this case, ya, mysql can run on a lower memory footprint than oracle.

> on the other hand, on my firewall (486, 32MB ram) MySQL can handle Apache logs (only about 200k hits, though) taking a split second for any reasonable query.

ah, yeah, 1995-era equipment.

ok, now, how about a real-world situation. Try this:
    - assuming that your company firewall is getting 5,000,000 rows/day
    - query: "is the current accept & deny traffic for any given protocol and port 5-minute interval more than 3 standard deviations from the mean for its time of day and day of week over last 90 days?"

I write queries like that all the time on db2, and it handles them easily - even spanning hundreds of millions of rows. And this is on a single server, not a distributed one. And I could do the same on oracle. But mysql? Not a chance.

Re:Move along, move along ...

[KiloByte]

I guess my main point was (to put it into your analogy) if you're not an F1 driver, don't bitch because you don't have the skills to drive an F1 car.

It's not a matter of skills -- it's a matter of not needing a costly, gas-guzzling car when you have better alternatives. Skills are something you can get quite easily; while it takes years to be a top-notch DBA, an intelligent person with experience in related fields (programming, sysadminning, etc) can learn the basics in a week. The thing is, if you take the F1 for a test drive, and it takes X gallons of gas, you can say "if I train driving it, I can need only X/4 that much fuel, but that is still 10 times as much as a normal car takes", you will learn that the other choice is a lot better unless you specifically need a car that fast. I did evaluate Oracle, and learned that even if you divide its requirements by 4, MySQL can still run circles around it within the needs the company or any of our customers have for a database.

Even if you are a master at using rocket launchers, you still can't conceal one in your pocket.

Re:Move along, move along ...

[KiloByte]

The only validity to your claim is that mysql might be a better choice if you're running 1995-era hardware, or are buying new hardware but want less than 128 mbytes of memory (for some unfathomable reason). In this case, ya, mysql can run on a lower memory footprint than oracle.
[...]
ok, now, how about a real-world situation. Try this:
Try this: host 100 clients on a single server (+backups). 100*128MB vs 100*10MB... *bzzzt* Your solution simply doesn't scale.

Try this:
        - assuming that your company firewall is getting 5,000,000 rows/day
        - query: "is the current accept & deny traffic for any given protocol and port 5-minute interval more than 3 standard deviations from the mean for its time of day and day of week over last 90 days?"
Ugh, and you really do that in SQL? You're using an abysmal tool for this job. By using the right data structure, you can answer these queries in real time on any hardware, with a minimal memory footprint and O(1) access time. And while you can say that you have a modern server doing this task, your solution doesn't scale, while mine does.

Re:Move along, move along ...

[nettdata]

You should go back and re-read your original post.

I've tried to be nice, but you're a moron (or a troll), so I'm done wasting my time with you.

Re:Move along, move along ...

[Zathrus]

The problem with Oracle is, it doesn't scale at all. It is meant to do grid computing, but can't really do anything smaller.

Man... all those Oracle customers that aren't running grid/cluster must be just plain stupid then.

A minimal installation of Oracle 10g takes ~800MB of memory, and will take over ten hours to install on a machine with 512MB ram

You're full of it. A brand new install of 10g Enterprise Edition, accepting all the defaults, takes 256MB of memory on a 512MB system and takes nowhere near 10 hours to install (not even half of that). And I know this from personal experience, since I've installed 10G EE on several boxes here at work (and I keep telling my bosses that I'm not an admin and know nothing about being an Oracle admin, so I'm quite sure I could be more efficient about both the install footprint and time).

Oh, and you're telling me that MySQL takes up 8 MB of memory based on your own statements. I sincerely doubt that.

Is Oracle the right solution for your case? No, probably not. But that doesn't validate your statements.

Re:Move along, move along ...

[kpharmer]

> Try this: host 100 clients on a single server (+backups). 100*128MB vs 100*10MB... *bzzzt* Your solution simply doesn't scale.

Sorry, I really don't think I know of 100 trivial databases that will fit on your 486. I'm sure there are probably tons of little team bowling score databases, grandma recipe databases, etc. But I just can't put my finger on more than a couple.

I don't work on oracle on daily basis, but I do work on db2 which is similar (at least when compared to mysql). And I can tell you that if you did want to put 100 tiny databases on a single db2 server with 1 gbyte of memory you could make it fit. I suspect you could do it almost as easily in Oracle.

Also, in the case of DB2 I suspect that you could probably put *more* databases on db2 for the same amount of memory than mysql. I don't know this to be a fact, but I do know that db2 offers you far more memory & database controls than mysql does.

>> Try this:
>> - assuming that your company firewall is getting 5,000,000 rows/day
>> - query: "is the current accept & deny traffic for any given protocol and port 5-minute interval more than 3 standard deviations from the mean for its time of day and day of week over
>>last 90 days?"
>Ugh, and you really do that in SQL? You're using an abysmal tool for this job. By using the right data
> structure, you can answer these queries in real time on any hardware, with a minimal memory footprint
> and O(1) access time. And while you can say that you have a modern server doing this task, your
> solution doesn't scale, while mine does.

The difference is that I'm not targeting discarded computers, but I am targeting low-labor. Hardware is cheap compared to labor. And that query query can be written and promoted into production in 1 hour against an existing database. Meanwhile, you'll spend another $10k in labor to create a real-time monitor that'll probably have to be reset every time your box is rebooted - all to save $5k on hardware.

Databases have been used for these purposes for ten years now. Unfortunately, your understanding of their capabilities may have been limited by the use of a limited tool incapable of this kind of (increasingly common) analytical processing.

Re:Move along, move along ...

[Anonymous Coward]

To the idiot who modded this as flamebait. Not only Eivind is a very experienced FreeBSD developer (hint: he knows what he's talking about), but everyone who has worked with a real RDBMS like PostgreSQL or Oracle can easily tell that all MySQL versions prior to 5.0 are a piece of shit when it comes to features and reliability. When I saw the new features in 5.0 I almost fell off my chair. Real databases have had views and stored procedures for years. Stop blindly advocating MySQL.

Re:Move along, move along ...

[bampot]

A minimal installation of Oracle 10g takes ~800MB of memory, and will take over ten hours to install on a machine with 512MB ram

Absolute garbage. Granted, the minimum recommended memory is 512MB, however I have 10g Enterprise running on an old P4 1.6GHz running CentOS with 256MB RAM. It isn't fast, but it still works. And it took about an hour to install. On my other machine (Athlon XP1800, Windows XP) with 1GB RAM I have two instances running, and you wouldn't even know it was there.

This isn't a hobby, I've been developing Oracle DB applications for 12 years, now I'm studying for the OCA/OCP DBA certification.

Granted, if you only have 17MB of data, Oracle is overkill but nowhere nearly as memory hungry as you make out. Check your facts first and stop picking numbers out your butt.

Re:Move along, move along ...

[zardo]

You really oversimplify things. MySQL isn't faster for some things, I've noticed writes with MySQL are extremely slow, and the query optimizer isn't as good as Oracle. I have become an expert at manually optimizing queries for MySQL, but when I use other databases, Oracle and Sybase mainly, the queries are optimized for me automatically.

I guess you just need to be doing more complex queries and transactions to really appreciate the big commercial databases.

I don't think I've ever created, used or otherwise been a part of any project where the PRODUCTIOn databases required more than 4 gigabytes. I have done developmental things that required tons of space, building production ready tables from a bunch of data I didn't really need.

I'd say the only thing you wouldn't want to use this limited version of Oracle for is some sort of full-text search engine. MySQL supports that and I can see tables growing over 4 gigs if you were searching full pages of text. However, many of the databases I've worked with are on dual-processor machines, with 4 gigs of ram (that seems to be a pretty common configuration for me right now), I don't know if all that ram is ever really used though. I figure in most cases I probably don't use a gig of ram, especially if your data is under 4 gigabytes. You might end up with 200-300 megs of indexes on 4 gigabytes of data.

There is also the crippled version of Sybase, which isn't AS crippled as this Oracle, they're comparable, Oracle may be a bit more advanced, judging from benchmarks I've seen.

Re:Move along, move along ...

[Eivind Eklund]

There's a few reasons we use MySQL:

(1) Migration costs; we've got significant codebases that are somewhat tied to MySQL, without automated tests for most of them, making a migration risky. We've also got some fairly large sites running on it; different performance characteristics could create a lot of issues.

(2) Emotional investment by my coworker.

(3) It's supposedly slightly easier to do various production hacks (copying tables etc) with MySQL than with PostgreSQL, and my coworker insist these are important. I've never ever ended up using these, instead of using dumps and hot copies and logging etc, so I personally can't see the need for these.

For my own evaluation, (1) is important, though (2) and (3) are what's really blocked the migration - if I'd been able to do all the choices without fighting over them, I'd have started a slow migration a year ago and been over on PostgreSQL now.

Eivind.

They must own stock in Maxtor

[Anonymous Coward]

Clearly, Oracle has bought a major disk drive company. Have you ever downloaded or tried to install Oracle? It's easily 10 Gigs of useless crud, wrapped around a few CD's of material actually relevant to your particular setup. For Linux, they publish it as a set of binary bundles that have to be strung together so that you can *then* take apart the tarball. What a waste of disk space!

The approach shows up in everything they do. Build a huge, conglomerated edifice of software to provide the one brick you actually need, rather than keeping components modular and portable. It's like making people install a whole radio station just to get a pair of headphones.

Re:They must own stock in Maxtor

[Anonymous Coward]

Uhm, no. It isn't that big at all. About 2 gig for the db server.

Re:They must own stock in Maxtor

[Ledis]

What are you talking about? Oracle 10g Linux is 638MB in size. And I have downloaded and installed it many many times on five different platforms.

Re:They must own stock in Maxtor

[quantum bit]

Uhm, no. It isn't that big at all. About 2 gig for the db server.

2 gig?!?!?! For a DB server?! Not big at all?!!!!!

$ ls -l /usr/ports/distfiles/postgresql/

-rw-r--r-- 1 root wheel 7825300 Oct 3 20:25 postgresql-base-8.0.4.tar.bz2
-rw-r--r-- 1 root wheel 2227623 Oct 3 20:26 postgresql-docs-8.0.4.tar.bz2
-rw-r--r-- 1 root wheel 389944 Apr 10 2005 postgresql-jdbc-8.0-311.src.tar.gz
-rw-r--r-- 1 root wheel 133881 Oct 3 20:26 postgresql-opt-8.0.4.tar.bz2

Re:They must own stock in Maxtor

[argent]

638 megabytes for a non-GUI app?

What the hell is in there?

How much does it unpack to?

Re:They must own stock in Maxtor

[cerberusss]

There are a few preconfigured database images on the CD, that's what's taking the space.

Re:They must own stock in Maxtor

[argent]

Errr...

How big is a "preconfigured database image" then?

Or have they pre-loaded these database images with porn or high-quality video of Larry Ellison or something?

How big is it without them?

Re:They must own stock in Maxtor

[cerberusss]

There's a bunch of testdata in there, including binary stuff to horse around with BLOBs and CLOBs. Don't forget, an Oracle database contains so much more than the data it stores. Each and every setting is in there, along with the data dictionary.

Re:They must own stock in Maxtor

[argent]

Ya didn't answer the question.

So how much space is that? How much is actually Oracle?

Re:They must own stock in Maxtor

[cerberusss]

I've looked it up:

• the included starter database is (packed) actually 100M.
• It's not a non-GUI application anymore: it comes with a Java- and webbased management tool which is (packed) also 100M.
• Also included is a Java SDK, a Java runtime and a complete Perl installation...

Re:They must own stock in Maxtor

[argent]

... leaving how much for Oracle? 400M? 300M?

Re:They must own stock in Maxtor

[cerberusss]

It's kinda difficult to see, because there are so many packages on that CD. But what I've done is the following: $ cd database/stage/Components
$ du -ms *rdbms* *plsql* | grep -v seeddb | awk '{sum=sum+$1;} END {print(sum)}'
90
$
The total size of all packages with the word 'rdbms' or 'plsql', but excluding the seed database. This adds up to almost 90Mb, packed.

Nice

[Delifisek]

And I'm not sure free Oracle better than mysql for Lamp (laop ?) project.

Oracle requires lots of attention to work. (O course if nobody touces it will work for ages) Any misuse may halt entire db.

Anyhow, I'm not sure that kind of movements stops even slows MySql and PostgreSQL.

And this kind of movements shows us FEAR...

Even DB giant Oracle was fear from MySql and Posgre SQL

Well done boys well done...

Re:Nice

[1001011010110101]

Any misuse may halt entire db.

Lol Such as?

Re:Nice

[Anonymous Coward]

There is one obvious use for it. Have any of you dealt with the Siebel bug-tracking system, popular in large companies? Siebel runs on top of an Oracle databse, and the clients for it are quite expensive and were written by crack monkeys. Access to any Siebel ticket means you can't open another Siebel window at the same time, anything you type in a Siebel window gets committed without the use of a commit button, so if you make a typo and hit "back" you've just screwed up the ticket, etc., etc.

If you download the Siebel database to another Oracle database and slap a decent web interface on the front end, you get free, usable web clients that can access multiple tickets, provide user information on tickets they may have outstanding, and allow users to see what the state of their trouble tickets are or look up the state of similar tickets to see if a problem is already supported. This is a huge advantage for people forced to call a helpdesk to look up the status of tickets, or for helpdesk workers who want to reference one Siebel ticket in another related ticket. And because it's a relatively small database of relevant information, it should fit quite comfortably inside the crippleware limitations of the free Oracle.

Re:Nice

[Anonymous Coward]

1. MS used to offer MSDE as a free database limited to 1Gb
      for a number of years already. It utilized exactly the
      same data format as full SQL Server. So a database backup
      in one could be restored to another 100%.

2. If the database is designed properly, transaction log
      shrunk weekly and most duplicate strings are put in a
      reference table and used by ID, most small businesses
      with about 100 real customer transactions per day will
      not need more than 100-200Mb of data per year. Thus
      either of the free versions will work just fine for a
      number of years.

3. When the time comes to scale up - realistically the
      business would be making a lot of money already. The
      $1-2K difference will be less than an opportunity cost
      of a lost day of sales/services.

I don't know what the licensing for real MS SQL Server
is, but assuming about $2-4K for about 10 connections
I see it still compete very heavily against Oracle at
$160/user ~= $1,600 for a similar level license.

Someone posted once that Slashdot is running on MySQL.
How large is Slashdot's database? and at what rate
is it growing?

Re:Nice

[mysticwhiskey]

Maybe "DELETE FROM DBA_OBJECTS"? Of course, if that's a "mishap" I'd hate to see how such a development company really does business... ;)

Re:Nice

[1001011010110101]

Yes, and probably format C: (where C: is where the datafiles are) might also trash the database. Same as your case, I woudn't consider that a "misshap" that can crash the whole engine as the OP implied. I was wondering what was he thinking about...

I've worked a bunch with Oracle, and never found something like he implies that woudn't happen in any database. In fact, the crash recovery features in Oracle are pretty good, and I doubt the mysql or postgress include something as advanced.

I use SQL databases for everything.

[bigtallmofo]

I can't imagine storing relational data in anything other than a server-based SQL database at this point. I have a hammer and everything looks like a nail to me.

Since I work all day every day with SQL Server databases, it's more of a nuisance for me with non-commercial projects to go to Access or similar non-server-based offerings.

Weak passwords?

[bigtallmofo]

Will the free version of Oracle be subject ot the same weak Oracle password encryption scheme that the commercial version is?

I've duplicated a number of techniques in the SANS article to make me leery of password security on my Oracle machines.

Re:Weak passwords?

[j_snare]

Yes, I imagine that the free version will have the "same weak Oracle password encryption scheme" that requires admin access (or at least select on a restricted table that wouldn't be granted in the first place unless the DBA wants a security hole).

What is people's problem? Yes, security is weaker in this than in some other systems, due to a hole that doesn't exist until someone with admin privilges explicitly grants it. Oracle actually admits to the security risks of granting privs on dictionary tables/views or to certain roles that have those privs. They suggest that you don't use them by default, but create what you need.

I duplicated several of the techniques that they describe for actual use in our own system last year (under tightly controlled conditions). Yes, they are security risks, but by default, they are controlled. Oracle will let you shoot yourself in the foot, but so will most other systems.

I'll be happy to burn Oracle on their installer, OEM, any other interface, install documentation, and even support (Metalink's running at being able to answer about 50% of my questions so far for the last several years). But the security bit in that CNet article is purely alarmist, and shouldn't have surprised or alarmed any competant Oracle DBA.

But will it be easy to install?

[samuel4242]

The main reason I like MySQL is it works five minutes after I finish downloading it. And it's much smaller than Oracle so I can download it quickly. I spent two days trying to make Oracle work on an Linux box and it never did. The price ain't the only reason I like open source. :-)

Re:But will it be easy to install?

[Anonymous Coward]

5 minutes? It took me 5 minutes to make it work the first time I set it up! (which was less than a week ago, so maybe I missed something? hmm, better check)

Re:But will it be easy to install?

[Eivind Eklund]

MySQL doesn't work. It is up and running five minutes; significant difference.

Yes, I work day to day with MySQL, due to the cost of migration. No, I don't feel good about routinely having to deal with corruption in production databases. It's not acceptable. Yes, it also happens with InnoDB.

Eivind.

Re:But will it be easy to install?

[Anonymous Coward]

It only takes 5 mins to install access, and it "works" straight up. Big F*K'n Deal. Just coz you call it a database don't make it so.

Re:But will it be easy to install?

[Anonymous Coward]

> I spent two days trying to make Oracle work on an Linux box and it never did.

Well, so much for your admin skills then ;-).

Re:But will it be easy to install?

[zardo]

Watch for the free Oracle edition ebuild script.

I hate Oracle with a passion

[Anonymous Coward]

Just wanted to share! They really need to work on the developer experience. I'm no MS apologist, but SQL Server is an absolute dream by comparison - and there'e a free version too.

Re:I hate Oracle with a passion

[BarryNorton]

I'm no MS apologist, but SQL Server is an absolute dream by comparison - and there'e a free version too.

I'd have modded you up if you'd named and linked it, but as it is I'll post: http://www.microsoft.com/sql/msde/default.mspx

Re:I hate Oracle with a passion

[kabz]

What people really mean when they say this is that SQL*NET is a POS.

SQL Query Analyser on MS SQLServer is a decent app, but any quality (read $$) third party software such as SQL Query Navigator or TOAD will make using ORACLE a whole lot more pleasant.

I work on massive queries that pretty much push both SQL Server and Oracle to their limts, and both databases can take it, though it feels like Oracle ultimately has more capability. Certainly, when you are writing complex software such as corporate billing with 30 - 40 connections running simultaneously, both apps are pretty good.

Thanks to the person earlier who said Oracle is available for Mac OS X. I'm downloading now. I really need to try and boost up my familiarity with SQL*NET. Spit, cough!!

OK support in Visual Studio

[spectrokid]

Actually their new plug in for Visual Studio is not bad at all. Personally I think it is better then JDevelloper, provided you don't code Java, of course. The main thing I have against Oracle is that I can't make e.g. an Access file which works out of the box on every computer. You always need to get the drivers set up properly first, and you can't do that without (local) administrator rights. That is where MS (obviously) can shamelessly profit from their monopoly.

Of course they will..

[Lord Bitman]

now that we already signed up for partnerworld to get the developer versions

And you thought Bill was the Prince of Darkness

[flipper65]

By all means, dive into Oracle, it will be a pleasure to see the Ellison flames replace the Gates flames.

I am choosing Oracle over MySQL

[backslashdot]

I am choosing Oracle over MySQL. Why? Because SCO is still refusing to sheild MySQL users from lawsuits that directly result from this deal with SCO. Note: I am not asking to be protected from whacko frivolous lawsuits, I am asking that MySQL protect us from a lawsuit by SCO that results from SCO making the allegation that MySQL placed commercial SCO code in MySQL in violation of _this_ agreement.

If MySQL is super confident that this deal cannot possibly result in any lawsuit to us, they can easily indemnify us from SCO lawsuits.

According to SCO's press release ( http://ir.sco.com/ReleaseDetail.cfm?ReleaseID=1720 37 ):

"The SCO Group, Inc. ("SCO") (Nasdaq: SCOX), a leading provider of UNIX(R) software technology for distributed, embedded and network-based systems, today announced that it has entered into an agreement with MySQL AB to jointly deliver a certified, COMMERCIAL version of the popular MySQL database for SCO OpenServer 6"

Clearly there is wiggle room for them to file a delusional lawsuit styled after the IBM lawsuit claiming that MySQL used SCO resources and knowledge to "improve the GPL version". Stop laughing, this is what they are saying about linux. And when that goes down in flames, they'll need another whacko lawsuit to pump their stock up one last desperate time so the execs can completely sell off their remaining shares.

Once more, if MySQL is super confident that this deal cannot possibly result in any lawsuit to us, they can should have no problem sheilding us from SCO lawsuits.

Re:I am choosing Oracle over MySQL

[Antique Geekmeister]

No one can indemnify you from delusional lawsuits, unless they have the spare money to fend it off or to succeed well enough to take back their court costs from the plaintiff. MySQL developers just don't have that kind of spare money.

Also note, SCO's lawsuit is draining their corporate sponsorship from Microsoft dry, and it's pretty clear from their income statements that they are being sponsored by Microsoft. Take a look over at http://groklaw.net/ for details. They don't have the resources to file another frivolous stock-pumping lawsuit like this one after they lose to IBM, becuase they're spending their core resources to pursue it. So if you're scared of SCO, I don't think you should worry about them after this: they haven't generated genuinely new or useful products in years.

Re:I am choosing Oracle over MySQL

[backslashdot]

No one can indemnify you from delusional lawsuits, unless they have the spare money to fend it off or to succeed well enough to take back their court costs from the plaintiff. MySQL developers just don't have that kind of spare money.

That's fine. I understand that. Then don't enter into agreements (especially with known scum with a dark history of suing _users_ of GPL software).

If I can't afford my rent, I dont go gambling with other people's credit cards in known swindling casinos.

Everybody's doing it

[Dekortage]

So MySQL and PostgreSQL have been free... then IBM announces a free version of DB2... then Microsoft says it's going to release SQL Server Express for free. So Oracle is playing catch-up. I wouldn't expect a major migration from MySQL to anything else; the conversion costs would be too high. But in the future, choice is a good thing.

Re:Everybody's doing it

[ergo98]

then Microsoft says it's going to release SQL Server Express for free

SQL Server Express is just the 2005 update of the MSDE, which has been around since SQL Server 7.

Of course Microsoft wasn't trying to fight open source - they were primarily trying to get developers to use SQL Server (and derivatives) as the data layer of their apps, instead of the traditional MS Access/Jet that was prevalent. Before it most developers would respond "I'm not making my customers pay for SQL Server", so Microsoft made the MSDE. They knew that many customers would eventually grow into the "real" SQL Server.

Re:Everybody's doing it

[karearea]

I think Sybase has done that as well - just quickly though I think it is only for Linux, and can't be used for a production environment.

--------------------
I was wrong once before, but that was only because I thought I was worng.

Oracle bought out InnoDB, now this.

[The Slashdotted]

Are they starting an arms race against MySQL?

Export Controls

[Anonymous Coward]

Did anyone check out the "export controls" for the download?

Good for a laugh if nothing else:
(http://www.oracle.com/technology/software/htdocs/ xe_lic.html?http://www.oracle.com/technology/softw are/products/database/xe/htdocs/102xewinsoft.html)

-I am not a citizen, national, or resident of, and am not under control of, the government of Cuba, Iran, Sudan, Iraq, Libya, North Korea, Syria, nor any country to which the United States has prohibited export.
-I will not download or otherwise export or re-export the Software, directly or indirectly, to the above mentioned countries nor to citizens, nationals or residents of those countries.
-I am not listed on the United States Department of Treasury lists of Specially Designated Nationals, Specially Designated Terrorists, and Specially Designated Narcotic Traffickers, nor am I listed on the United States Department of Commerce Table of Denial Orders.
-I will not download or otherwise export or re-export the Software, directly or indirectly, to persons on the above mentioned lists.
-I will not use the Software for, and will not allow the Software to be used for, any purposes prohibited by United States law, including, without limitation, for the development, design, manufacture or production of nuclear, chemical or biological weapons of mass destruction.

Granted I'm not familiar with Oracle, but I don't think it'd be too handy in the manufacture of chemical or biological weapons.

Re:Export Controls

[badfish99]

This is just symptomatic of the bureaucratic fantasy that you can do "security" by asking people questions. It's like the "did you pack your own suitcase" question at the airport. Of course everyone is going to say "yes" even if they didn't, but the people asking the question haven't worked that out.

Re:Export Controls

[Anonymous Coward]

Or, as the situation requires, say "No, my wife did..."

Re:Export Controls

[Anonymous Coward]

Can ANYONE truthfully respond to this?

-I am not listed on the United States Department of Treasury lists of Specially Designated Nationals, Specially Designated Terrorists, and Specially Designated Narcotic Traffickers, nor am I listed on the United States Department of Commerce Table of Denial Orders.
-I will not download or otherwise export or re-export the Software, directly or indirectly, to persons on the above mentioned lists.

Are the lists available for public inspection? Any one of us could be on those lists, unless you believe that they're 100% accurate. And how can you know when you pass a CD to a co-worker that THEY are not on one of the lists?

Even without a tinfoil hat checkbox the questions are clearly stupid.

Re:Export Controls

[Herschel Cohen]

On the Linux download side, the license was explicitly limited to a beta version with a cut off date along with the option of an announced revocation by either party. If that condition is just not legal boiler plate a free, limited version of the Oracle database has a short life expectancy.

Re:Export Controls

[Decker-Mage]

The Windows side has the same cutoff, 31 Jan 2006. That's fairly standard for a beta I might add. However, this says nothing about what, if any, they will charge for the Express Edition. Is it just me, or did anyone else actually read all the terms of their license agreement and find them just a bit over the top? Not even MS goes to these extremes and I read all agreements. [The community agreement was even odder but I won't go there.]

I'll be downloading it later after I finish collection SQL Server 2005 Enterprise here ;-). It'll be interesting to see how well it works with my Oracle speaking tools here.I do hope they work as I really, really don't want to face yet another tool upgrade cycle. Those are expensive!

Re:Export Controls

[samuel4242]

Mod this up! How do we know whether we're on the list?????

Re:Export Controls

[Herschel Cohen]

I share your wonder about others reading carefully, however, not just limited to the license. I am noticing many stories (not just on /.) where the summaries are inconsistent with the cited article content and/or linked information that is contradictory to the poster's thrust.

[I was tempted too to wonder about why were Oracle's terms deemed superior to MS routine offers of 90 day test version of SQL Server. For me there are several turn offs to the latter option, but here many users might opt for its try out terms.]

Re:Export Controls

[Decker-Mage]

Sooo, I'm not the only one clicking on article links to insure that what the article asserts bears at least some resemblence to the original source. If it were only the WWW, I'd rely on print media but it seems print articles also have the same problem especially when it comes to scientific/technological issues which is extremely sad. There's not much either of us can do about it save point out the error when we come across it but so far I've yet to see one of my letters make it into print which is about what I expected. I actually do better on the WWW where it seems that having alternative avenues encourages a certain amount of wary openness. Whatever. That's a debate for another day/article ;-).

Re: SQL Server 2005. MS is not withdrawing the CTP versions which will still remain available for download after the Nov. 7th release date according to the most recent MSDN Flash that I received yesterday (actually several times so I guess they wanted to make a point!). That version is good for over a year before it times out and you can go all the way up to the enterprise version which has some adequate built-in tools (they ain't Embarcadero ;-). So, if anyone out there is interested, you may want to grab a copy although they are rather hefty in size (Enterprise weighs in at 920 MB). Still, you can use those tools with the Express version or take what you've built and use it as there isn't that much missing between the versions for the personal/workgroup/small-business level which is the level I'm mostly working with when I'm not testing for the big-guys.

Just my $.02.

Compete with SQL Server and MySQL

[olddotter]

I have been thinking about this for a while now, since Oracle showed interest in Open source databases. I think the goal is to offer a barrier to attack moving up from the low end of the database market.

A few years ago a well placed Oracle employee gave a talk at the local LUG meeting. It was plain from his attitude that if you didn't have at least a 4 way box hosting your database he didn't see you as a customer or potential customer. Oracle appeared blind to the needs of small business. And of course that is where (IMHO anyway) MS SQL Server posses the biggest threat to Oracle.

I think Oracle will use MySQL and their own free offering as a way to compete with MS in the low end of the market.

Re:Compete with SQL Server and MySQL

[Herschel Cohen]

Here is a link where it is suggested Oracle might use MySQL to squeeze MS's SQL Server from above and below: http://lxer.com/module/newswire/view/45935/index.h tml

Smart move, but it could be a trap :)

[OpenSourced]

I think that's a smart move by Oracle. From a long time ago they have allowed full download of their databases for testing purposes. I have a copy of Oracle 9 running in my machine to make tests of software I develop. So my customers with Oracle have better service. Probably some copies of the database will end up as production units, but few companies will trust its bussiness data to an illegal piece of software.

So Oracle has realized that the free availability hasn't cut into their sales. The next step is logical. You give away an entry-level database (entry-level users would probably use an illegal copy, or worse, an open source db), and then wait till the needs grow and they need the real thing. If the needs don't grow, well, who need those little-bussiness-that-don't-grow as customers, anyway ?

I see the thing as mainly good for the users and developers. Of course it'll cut into Open Source databases, but they'll still have their niche. After all, you should be careful with what you do with this free Oracle. Oracle can change its mind in two year's time and leave you with all your data and processes in a database that won't be supported or upgraded anymore. You'd have fallen into Oracle's trap. That's much more difficult to happen with an Open Source database.

Re:Smart move, but it could be a trap :)

[Rorschach1]

"From a long time ago they have allowed full download of their databases for testing purposes. I have a copy of Oracle 9 running in my machine to make tests of software I develop."

This is one of the things I like best about Oracle. Yeah, installation isn't always easy (it's much better now than in Oracle 7) but at least they never make you jump through hoops when it comes to license management. No keys to enter, license files to install, or anything. I think it gives you a little warning in the installer to make sure you've got appropriate licenses for all options, but it never forces anything.

Re:Smart move, but it could be a trap :)

[UnanimousCoward]

I agree with the parent post, and I think that /.ers are missing the point. Oracle is not going after the build-it-and-run-MySQL folks. There are applications that work with Oracle that just don't have the support of the existing free databases. Oracle is giving companies a chance to try out these things.

Take for example, GIS spatial applications. You say that MySQL supports spatial? Ever try to mix in transactions? This is just one example. There are plenty of others. Now, Oracle has removed one price barrier--yeah, yeah, there are DBA costs to be considered, but now, a company can at least check of the licensing fees (which are NOT trivial) as one reason not to give Oracle a shot.

Oracle Licensing

[HaloZero]

I've used 8i and 9i in the past, without much thought or care as to how it got where I was using it from, or how costly it was to the respective company. Could someone from the trenches clue me in as to typical Oracle licensing practices? Like, if I want a box that can take 5 connections or 50,000.

Re:Oracle Licensing

[mzito]

Oracle is licensed by the processor or by "named user", not by simultaneous connections. For standard edition, its $15k per processor, or $300 per database user (plus a minimum of 5 users). For enterprise edition its $40k per processor or $800 per named user, with a minimum of 25 named users per processor in the system. This is before you add any of the expensive options like RAC, Partioning, etc. that can add $20k each to the price.

There's also standard edition one, which is cheaper than either and supports some of the advanced features of both. It's designed to compete with some of the SQL server shops that have HA requirements but aren't willing to pay for oracle enterprise edition.

Of course, all of these prices are list, and for good negotiators, discounts upwards of 50% off list are not uncommon.

Thanks,
Matt

Re:Oracle Licensing

[HaloZero]

Holy hell, that's outrageous. The fact that they can pull 50% off list for negotiators just shows how much profit they're making off of the project. I'll stick with MySQL... or CowboyNeal with postits and crayons.

Re:Oracle Licensing

[will_die]

Oracle is very open in its pricing. check out http://oraclestore.oracle.com/ and you can customize what you want.

Re:Oracle Licensing

[bobaferret]

Believe it or not, this isn't outrageous. The software that my company writes sells for 30-40K for a license and 3-4k per year for maintanance. You throw in the as400 that this has to run on, and you've quickly hit 100K. Add about 10K per extra software package btw. It's not uncommon for new clients to drop 70K in software costs. However, are nearest competitor, not in sales but price, starts at 300K and quickly get's into the millions. I had to adjust when shifting out of an idealistic academic atmosphere, into the biz world. You have to remeber that most software shops couldn't run without charging prices like this. You also have to remember that there are two kinds of software shops out there. Retail and Business. Retail works on selling lots of coppies for a cheap price while the biz market works on selling a few copies for a very large price. Oracle tries to serve the biz market and not the retail market, and there prices are quite inline, if not cheap for that market.

BTW, we use postgres for everything not running on a as400..

Re:Oracle Licensing

[mzito]

It gets even crazier, believe it or not. So, what exactly is a "named user" - I've run into customers that have been told by Oracle that a named user is a consumer of their application - so if there are 100,000 people using the web app, you need 100,000 named user licenses.

This leads to per-processor. If you have three databases running on an 8 processor server, you need 24 processor licenses.

On the flip side, Oracle has some really amazing functionality and stability. Is Oracle needed for every application/worth it for every application? No. Is it needed for some? Sure.

Thanks,
Matt

Re:Oracle Licensing

[Anonymous Coward]

I'm a database consultant - I have a good deal of experience with many databases, mostly Oracle. I've been using Oracle since 7.3 (about 10 years), and I would say I know about 10% of it's features.

Oracle is really NOT that expensive, when you consider that staffing (2-3 DBAs) will be at least $300K/year. Machines and maintenance (test, dev, qa, prod, sandbox) at LEAST another 100K (even Linux boxes). Most people are using packaged applications on the database, and in reality they have only a few database choices (Sybase, SQLServer, DB2 and Oracle). If you app vendor doesn't support it, you are not going to use it.

A lot of the features, if you use them, can easily save you enough to make up the cost. For example, using DataGuard to do replication is much easier (and cheaper) than say SRDF on an EMC DMX.

People say Oracle is difficult to configure. This is true (somewhat). Out of the box, the init.ora is really not appropriate to anything. But this is actually good, as it forces someone to understand the application an configure it appropriately. Enterprise Edition is more expensive, but other than > 4 CPU support, almost no one uses the EE features.

Stored Outlines, histograms, materialized views, and (as of 10g) regular expression queries are great features that few people really use.

SQLtrace, statspack, and explain plan are great tuning features that I couldn't live without, but few people use effectively.

The Oracle documentation is incredibly good - better than any software product I have used in my life. 90% of the time I can find an answer to my question in the documentation.

All that being said, it would not suprise me if Oracle gives away their database in the next 5-10 years. (maybe not open source, but some free-ish license). They make their money from support contracts, fin apps, and consulting. They already own so much of the database market that there is no growth left there (at least profit wise).

Remember, database problems are old problems. In the last 40 years, they have mostly been solved. Yet people love to re-invent the wheel.

Re:Oracle Licensing

[briansmith]

Oracle is really NOT that expensive, when you consider that staffing (2-3 DBAs) will be at least $300K/year.

I was the DBA for an departmental Oracle database for a few years and I spent way less than a day a week on DBA duties. If you follow the SAME (stripe and mirror everything) methodology, buy enough memory, turn on automatic extent handling, and write a few scripts, you don't have to do much work, in my experience.

This is especially true if you have good developers that are capable of creating reasonable schemas without hand-holding from the DBA. If your developers know nothing about SQL then your DBA will have to do a lot of work, but it would be more productive to just teach your developers good database design.

Re:Oracle Licensing

[the_arrow]

Oracle is licensed by the processor or by "named user", not by simultaneous connections.

Heh, a sales representative for Oracle told the company I work for that each connection was a "named user". We replied that then we would re-code our applications to only use one connection, total, for all programs.
She wasn't happy having us as customers.

It's already available!!!

[alsutton]

You can download the beta for Linux and Windows from http://www.oracle.com/technology/software/products /database/xe/index.html

Re:It's already available!!!

[Norfair]

Direct link -> http://download.oracle.com/otn/nt/oracle10g/xe/Ora cleXE.exe/

MS has always offered free SQL Server

[The Slashdotted]

They gave it the stupid name for it's free SQL, MSDE, or Microsoft SQL Server 2000 Desktop Engine. It has no UI, everything is done by vendor tools and CLI.
http://www.microsoft.com/sql/msde/default.mspx

Re:MS has always offered free SQL Server

[LLuthor]

MS Access is "supposed" to be the interface to it. Access is actually a very nice interface and handles all of the SQL Server (2000) features extremely well, from simple views to complex stored procedures and macros.

I haven't yet had a chance to play around with SQL Server 2005, but I understand that the entire .NET framework and runtime has been deeply integrated with it, and as such, all .NET languages can be used to create first class database objects like stored procedures, and even custom data types.

Re:MS has always offered free SQL Server

[Anonymous Coward]

I understand that the entire .NET framework and runtime has been deeply integrated with it, and as such, all .NET languages can be used to create first class database objects like stored procedures, and even custom data types.

Yes and no. Yes, you can write .NET stored procedures. No, they have to fall back to SQL for database access, so they're not a complete replacement.

Re:MS has always offered free SQL Server

[TheSync]

It isn't too hard to find the Microsoft SQL client GUI tools to work with MSDE...they are available in books on Microsoft SQL, and also there are third-party GUIs available as well.

Re:MS has always offered free SQL Server

[killjoe]

Nice to know that after five years or so they are catching up to where oracle and java was. Why does it take them so long anyway?

Re:MS has always offered free SQL Server

[LLuthor]

Because they don't charge $50000 per CPU per year, and thus can afford fewer people working full-time on the code.

Even with full support, there are plenty of companies who will support it and the competition keeps the prices low... unlike Oracles.

Oracle is great if you really NEED that level of scalability, and those that do, already use Oracle or DB2 as no OSS solution meets their needs. But if you are not handling 100 million transactions per hour or working with terabyte datasets, then PG is just fine, and costs NOTHING.

Re:MS has always offered free SQL Server

[LLuthor]

Sorry, there should have some been sarcasm tags there, but /. ate them (dont flame me) :)

Re:MS has always offered free SQL Server

[killjoe]

"Because they don't charge $50000 per CPU per year, and thus can afford fewer people working full-time on the code."

Feature for feature oracle costs the same as SQL server. To think that Oracle has more money to throw at a project then MS is laughable.

Re:MS has always offered free SQL Server

[LLuthor]

See my reply my post. I was making a joke... slashdot happened to eat the sarcasm tags around it :)

Re:MS has always offered free SQL Server

[mgbastard]

I understand that the entire .NET framework and runtime has been deeply integrated with it, and as such, all .NET languages can be used to create first class database objects like stored procedures, and even custom data types.

Yes and no. Yes, you can write .NET stored procedures. No, they have to fall back to SQL for database access, so they're not a complete replacement.

Well what do you expect? Something like Enterprise Data Objects from NeXT? Apple's CoreData?

Er, wait... well it takes time to rip something off THAT good without the source code. Microsoft can't even get the fabled WinFS shipping after like a decade of trying.

SQL is so damn mature for data querying; it's still one of the best thought out languages we have for computing, IMHO. That being said, for some tasks, its overkill. And some neat stuff has been developed to overcome that. OO databases are a totally different beast than rowscanning, indexing, optimizing, queryplanning rocket science of a database server. There's some neat work in object-oriented interfaces from Apple and NeXT for your traditional database server.

Why Bother?

[yancey]

Sounds to me like Oracle's market share is dropping and they're doing what they can to hold on. Besides, they make the bulk of their money from support fees, not the database itself. If you do anything serious with their software, you'll NEED that support. It's been my experience that Oracle is so much more difficult to configure and optimize that I'll just stick with MySQL and other database products when needed. I am completely convinced that unless you require a feature only present in Oracle or unless you have software that requires it, you should go with another database.

Re:Why Bother?

[will_die]

IDC's report, "Worldwide RDBMS 2004 Vendor Shares: Preliminary Results for the Top 5 Vendors Show a Solid Boost (March 2005, IDC #32969)," estimates that Oracle(r) Database grew 14.5 percent year over year and increased its market share lead to 41.3 percent. Overall, the worldwide market for relational database management systems grew by 11.6 percent in 2004. In 2003, Oracle Database posted 8.6 percent growth year over year and was the market share leader with 39.8 percent.

Will Separate FOSS Fans From Freebie Fans

[reallocate]

Offering gratis but capable versions of closed proprietary software may or may not turn out to be a good marketing move, but it would certainly separate those who like FOSS from those who like freebies. Given the fact that the vast majority of FOSS users have no interest in modifiying source code, or the capability, I suspect most of us fall into the freebie camp.

Don't waste your time(stick with LAMP, stuff LAOP)

[Anonymous Coward]

It would be better spent getting stuff done with MySQL. Oracle have realised their market is on the slide. Students/Developers don't want to install a massive piece of bloatware, also Oracle has been smelling like a rotten corpse for a while now, it'll drop to the floor soon. It has taken them how long to come up with this? Pick a non-fly ridden solution. In a year or two they'll be trying to give away Oracle. Also LAOP, that's not as nice as LAMP.

Free?! Not cheap enough.

[ticklejw]

You couldn't pay me to use Oracle. I'm not sure what they expect to gain by this, but it certainly isn't my respect. If they want to generate more interest, they can make a product that works and doesn't require you to hire an Oracle-certified specialist to maintain all the little quirks and problems.

Go ahead, -1 flamebait or troll or whatever, I just really don't like their database software, and especially now that MySQL has the features that it does in version 5 and from what I hear PostgreSQL has been keeping up, there's no reason to pay half a million dollars for something sub-par. There's also no reason to pay free for the limited, restricted edition only to have to use your half million to upgrade when you outgrow the restrictions.

Yet another example of where Free software wins.

SCO after the flames

[The Cornishman]

When SCO vs IBM goes down in flames, maybe with Novell contributing its own flamethrower, there isn't going to be enough of SCO left to pump, just a few cinders blowing on the wind. At that point Redhat vs SCO gets decided too - I don't think SCO is short of things for their lawyers to do, to be honest.

Switch? Hell No!

[brennz]

I don't like the idea of switching to Oracle because it is not open source.

I can't sweet talk Oracle devs into including some new feature I want, not without going through loads of bureacracy. I can't submit patches to the Oracle code base. I have to worry about rampant security flaws. I have to pay way too much if my DB gets bigger. I have to put up with mediocre performance.

No thanks.

I am sticking with PostgreSQL. I can hop into #postgresql on irc.freenode.net and talk to bruce momjian about features and coding for postgresql. I can submit patches. I can review the PG codebase, and I know how fast the PG devs fix flaws. I don't have to pay anything if my DB gets bigger. I have great performance. Shoot, I'll even have multi-master replication (slony II) for free in the future.

Super!

[Lao-Tzu]

As a software developer in the petroleum industry, I find that all of our clients use Oracle for their database needs. The release of this product will allow us to test and tweak applications to work against Oracle without purchasing the database. This works out great for us. I don't want to use Oracle in any way, but now I have the capability to use it in the most minimal way that will allow me to sell and support software.

My prefered database system is PostgreSQL. It would seem that no level of marketing skill can convince anyone in this industry that Free Software has value. Funny...

Re:Super!

[nettdata]

You should know that anyone can download the fully un-crippled version of the Enterprise DB from http://otn.oracle.com/ and use it all you want for development without paying a dime in licensing fees.

It's only when you go Production that you're required to have your Oracle licenses purchased.

Re:Super!

[Lao-Tzu]

Thanks for the heads up. I wasn't aware of that.

The Rug

[StormReaver]

I won't be using any non-Free (note the capital F) database for either myself or my contract clients. Oracle, DB2, and Microsoft SQL Server may be free (note lowercase L) of charge for some limited range of uses, but none of them are Free as in Rug. You know the rug I'm talking about. It's the one that proprietary vendors like to pull out from under the people who get hooked on proprietary products.

Been there, done that, won't ever do it again.

As long it's not free-as-in-speech...

[vhogemann]

I'll not touch it.

Not trying to be a troll here, but why use Oracle when they won't support our Distro of choice (Debian)?

At work we have good experiences with Firebird, we have several databases, some over 1.6GB size, with more than 50 concurrent connections. And there was no downtime or corruption problems since the thing went to production, almost 3 years now.

Ok, Oracle has big advantages over Firebird. But they're worth moving away from Debian, a distro we trust and are confortable with? Are these advantages worth the extra money spent on licences for Oracle and it's supported Linux distros?

I work at a public institution, the healthcare department of Rio de Janeiro City, and there's barely enought money to run the hospitals, to buy medicine and such. Sure we could use this free Oracle, but we made such a long way until now using only OpenSource solutions. Why would we change now?

Just my 2c.

Re:As long it's not free-as-in-speech...

[xagon7]

I'll second that about Firebird. Damn fine database. With a running history of over 20 years, small footprint, capability of embeddedness, and extensions, and cross-platform capability, it offers WAY more than other Open Source databases. I just wish it had been opened up sooner. I hope more people will take a serious look at Firebird in the future.

Good news for Compiere

[jlebrech]

As compiere is based on the commercial version of 10g yet itself is supposed to be free, they can continue using the oracle db as the backend and still focus more on the actual system itself instead of wasting time porting to mySQL or postgreSQL.

The Munificent Oracle

[Anonymous Coward]

Wow! It looks like Oracle has turned over a new leaf, and is eager to join the free software revolution!

Not.

Think of "free" "light" versions of product as nothing more than an inadequate waste of your time. The purpose of these products is to get you up to speed on the product's featureset, so that when you figure out that the free version is inadequate, you'll buck up for the proprietary version, rather than backpedal and start all over.

This is a marketing campaign, not a newfound commitment to free software idealism.

Expect more of this..

[ahodgkinson]

Given the perceived popularity of MySQL, Oracle obviously feels it needs to react in order to prevent a slide in its market share. This is interesting, given that Oracle is one of the world's largest software companies and makes most of its money selling top-end systems to large enterprises, which isn't (yet) MySQL's playing field.

MySQL is good enough for many smaller software projects and is therefore capturing mind share in the developer world. Oracle obviously realizes this leads to a trickle up effect as software developers with MySQL experience will probably start to recommend it for other, larger, projects.

Oracle is trying counteract this by attempting to capture developer's mind share, rather than battling directly for market share. This is a long-term strategy and its success will depend on how well Oracle interacts and reacts with the Open Source developer community.

From the few comments posted here, mainly those stating how big and complex the Oracle system is, I wonder if Oracle actually gets it. If the learning (and administering) curve is really that steep, Oracle may be better off if it releases a light (in size and complexity) version that is easy to get up and running on small projects. A second recommendation would be to make sure Oracle 10 is included by default on most popular Linux distributions (which will be difficult, given it's size and complexity).

While I am impressed by Oracle's move, I'll be surprised if it gets them the gains they are hoping for. I don't think they realize the commitment this move will require in the Open Source world in order to be successful. Open Source is one of the few playing fields where actions still count more than PR.

This makes me wonder if another major software company will follow with a drastic reaction when the Linux desktop and the Open Office suite are truly ready for prime time.

We live in interesting times!

Re:Expect more of this..

[ameoba]

I have to wonder how the purchase of InnoDB is going to work into their strategy of fighting the creep of MySQL into the Oracle market. One thing that comes to mind would be to allow an 'easy' migration from MySQL to Oracle, perhaps even allowing MySQL tables to be used natively.

Re:Expect more of this..

[briansmith]

If the learning (and administering) curve is really that steep, Oracle may be better off if it releases a light (in size and complexity) version that is easy to get up and running on small projects.

You mean, like "Oracle Lite" that has been around forever, or "Oracle Express", which is the topic of this discussion?

Re:Expect more of this..

[zardo]

Come on everybody, Larry Elison just wants to be the richest man in the world just ONCE, let him have his moment! Come on!

Re:Expect more of this..

[Anonymous Coward]

Oracle Express is the old OLAP engine. The topic of this Discussion is XE.

Still has restrictions

[waif69]

If you look at the EULA you will see that this has a time restriction. "...provided to you by Oracle solely for evaluation purposes until January 31, 2006."

Obviously this is just a ploy to get developers to write apps on Oracle then, when the application has gotten fat, they will have to pay the fees for a version of Oracle that can support the app or rewrite the whole thing.

I think that only good reason to obtain 10g is to learn Oracle. If I was working at a company that was moving to Oracle, or at least talking about it, I would DL this to learn it for improved job opportunities.

Just my $.02.

Re:Still has restrictions

[kylegordon]

I could be wrong, but I think this time limit may be due to the fact that the current release is a beta. This way, they can effectively force people to move on up to the full fat edition when the time comes.

Re:Still has restrictions

[wandazulu]

The current version is a beta and the date refers to when the "shipping" product should become available, negating the use of the beta.

Three months isn't exactly enough time for apps to have gotten "fat". Besides, you don't have to use every aspect of the database (pl/sql comes to mind), but treat it as a straightforward mysql installation...stick with the dynamic sql and you really wouldn't know the difference.

Re:Still has restrictions

[Matt Perry]

I think that only good reason to obtain 10g is to learn Oracle. If I was working at a company that was moving to Oracle, or at least talking about it, I would DL this to learn it for improved job opportunities.

You can already download the full Oracle for free from Oracle's web site. Just create an account and then download. It's free to use for development and learning.

Re:Free?! Not cheap enough.

[Anonymous Coward]

PostgreSQL has more than been keeping up, buddy. It's way ahead of MySQL in all areas which matter to businesses such as ours (major ISP).

It is the de facto open source RDBMS out there. Nothing comes close.

I read recently that MySQL has finally come out with triggers/functions and row-level locking. Crikey, those elements have been standard in not only serious commercial RDBMS (oracel, sybase, db2, informix, etc), but in PG for friggin years...

Don't get me wrong, MySQL has a place, but it equates to MS Access in my eyes.

C'mon, people...

[Anonymous Coward]

If you want to catalog your CD collection, or serving a few thousand pages on your website a day, MySQL is a great choice. If you're running a .NET-driven website, use your SQL Server. These are all fine tools. Go in peace.

But don't complain about Oracle's reliability, features, scalability, or flexibility. When you need ass-kicking database performance and you really know what you're doing, those other tools are still in the stone age.

Though it's heresy in a world of open-source advocates (of which I am one, btw), the truth is the unpleasant fact that MySQL and Postgres are still relatively crude tools compared to Oracle. Compare Oracle's superior MVCC, hierarchical/analytic queries, flashback query, fine-tuned storage/space management, support for sophisticated indexes (reverse-key being one), "tune-ability," and on and on. There may be a learning curve, but seriously, the interface in 10g is pretty nonthreatening, so even the novice has a good shot of being able to do a lot of very cool things.

If you don't need the power, that's fine, don't use it, but don't bash something you dont' understand. Can I get a witnes?

Re:C'mon, people...

[sticks_us]

You know, this is actually a pretty good post.

While it's laudable to be rabidly open-source, the fact is that for larger, enterprise-style work, you're just going to need some heavy-hitting software, provided by a large vendor. I can see a couple of people sitting in a coffee shop debating the merits of postgres vs. DB2, but I can NOT see Deutsche Bank announcing that they're switching over to MySQL because of its suitability for that type of application. Just not going to happen, people. By the time some of the open-source dbs work out their kinks and add features, the major players are already another light year ahead.

Nothing to see here, move along.

SQL degrading towards end-of-life?

[Carl Rosenberger]

Some indicators that SQL does not have a bright future as a database interface for object-oriented programming languages:

LINQ/DLINQ
http://msdn.microsoft.com/netframework/future/linq /

Native Queries
http://www.odbms.org/
http://www.db4o.com/

What does Oracle have to offer in this direction?

Seeing the Gorilla drive the money out of the SQL stack is a strong sign that innovation is over.

Oracle is simply increasing knowledge supply...

[GiorgioG]

By introducing a free entry-level product, Oracle intends to get more developers and students familiar with its namesake database.

More like, intends to get more developers familiar with its namesake database to increase supply of Oracle-experienced staff and therefore lower cost.

SuSE Support

[Anonymous Coward]

Hopefully they'll support the OpenSUSE project at some point and/or the regular SuSE client... If not, guess I'm sticking with MySQL

Question

[CaptainZapp]

Could this be a badly disguised jab at the good burgers from Sybase, who offer their flagship product, slightly restricted, under the name Adaptive Server Enterprise Express Edition (cough) (Link takes you to the registration form) since almost a year?

In my opinion Oracle is one of the least trustworthy software vendors and I sure as hell wouldn't bank my company on them, regardless of the price they ask.

Re:Business plan

[yancey]

Just like drugs... the first one's free. Oracle makes their money from support contracts, not the products. If you're software is bad enough, everyone needs the support!

10g Express edition in response to MSDE

[Anonymous Coward]

Oracle are coming out with 10g Express edition in reponse to the long standing MSDE which is SQL 2000 desktop engine. The users can use it and redistribute it for free.
SQL 2005 is also shipping with SQL 2005 Express Edition.
I think Oracle's just looked at the success of limited functionality MSDE and decided that they want a piece too.

I am done now.

SQL Express

[MajorDick]

these requirments are nearly Identical to SQLExpress, the renamed MSDE from MS.

But on a windows system WOW is it handy for building apps with embeeded db , (1000 times better than Access, both in performance reliablity and coding for it.)
Hopefully Oragle will make it that easy for Unix/Linux?Solaris development.

What does this mean for [...] MySQL and PostgreSQL

[doedel]

Well, it doesn't affect them at all. And not Mr. Mendelsohn got that wrong but TFA and our /. poster.

You could download and test Oracle's DBs for free for quite a while - but only now you can use them for free in a production environment. That's clearly aimed at MS' offerings like MSDE and SQL Server Express and not at MySQL or PostgreSQL.

Also Oracle is an enterprise DB and MySQL, PostgreSQL or even MS SQL Server can not be compared to it in that regard. This also means you need trained staff to administer it - forget about just downloading and using it. Tried to get a demo of Oracle's XML Publisher working - I know what I'm talking about ;-).

So if your shop is already using Oracle's DBs this is a nice offer for the occasional small project. But for everyone else, just stay with what you know and love - whether it's MySQL, PostgreSQL, MS SQL Server or something else.

Compiere ERP

[dk.r*nger]

This could be a great boost to smaller compiere ERP solutions. This is a cool GPL ERP system, but it runs exclusively on Oracle. Different portings projects is in place, all very interesting, but not production quality yet. The catch is that a lot of the system is in stored procedures etc.
Convincing a business to cough up the high price for an Oracle license to use an unknown, yet brilliant ERP solution, is tough..
This will allow us to set up Compiere for a business for the price of the hardware and consulting (much easier), and then, when they hit the 4 gb data, 1 gb ram or 1 cpu limits, the price of an oracle license is more reasonable.
Yeah, it would be nicer if it ran off of PostgreSQL or Firebrid, but it doesn't.

Death of MySQL

[PhotoGuy]

Oracle is proceeding in a very smart fashion to eliminate MySQL, which I'm sure they see as their biggest threat. (Yes, for all your Oracle zealouts out there, Oracle can do this, and Oracle can do that, and the other thing, that MySQL can not; but for 99.9% of web based database-driven applications, MySQL works great and does everything you need.)

First, they buy up the InnoDB, which is the engine behind MySQL that offers the more advanced features (ones most likely to compete with Oracle), surely bringing an end to significant advances in the leading edge of MySQL.

Second, they offer a free alternative to get people onto Oracle.

This is very dangerous stuff for MySQL. Let's hope Oracle doesn't succeed in fully killing it.

Re:Death of MySQL

[Kent Recal]

I don't think Oracle are MySQL are in the same market.
As someone else cited, an Oracle license starts at around $15k.

Also you'll need at least one dedicated DBA to keep it ticking.

The pricetag alone reserves oracle for the "big guys" while MySQL is your average-joe database.
And feature-wise MySQL can't hold a candle to the big O either...

So, no, MySQL is not really threatening Oracle at all, right now.

too little.... too late.

[Anonymous Coward]

They are about 15 years late on this one. The only reason they do this is because they are about to fold under pressure from MySQL, Postgres, etc.

Business, People, Business!

[corcoranp]

Most import question any of us should be asking ourselves is how does this give Oracle a competitive advantage...or how does this make Oracle money? I have some ideas, but mostly just suppositions, not cold hard facts (Maybe I should read the earnings report I got from them?)

I'm a firm believer that there can be balance between OSS & Non-OSS (MySQL being a good example). It's troubling to see "Corporates" making certain power plays that in the end do not help the little guy. I suppose another question then is: Is this such a move?

Tester's heaven

[Craig Ringer]

Whether or not you'd like to use Oracle yourself, this is good news for software developers. It means they can deploy and test against a running version of Oracle with no need to worry about "developer program" memberships, trials that expire, and similar crud.

This'll be very helpful for me in ensuring that my code is portable across databases (at least PostgreSQL and Oracle).

This is an interesting development

[Anonymous Coward]

The company I work for has asked me to look at our future database options (we currently use MySQL 4.1 but are outgrowing it rapidly) and this product might open the door for Oracle. Ideally we'd love to use PostgreSQL but it doesn't seem to have the commerical support in the UK that we think we might need going forward.

IBM's DB2 is also a favourite - its a shame that these free "express" versions are just slightly below our requirements (or near future requirements). But it's funny are we are probably exactly the kind of small but growing company these Database Vendors are trying to attract.

I guess the point is (like other have said) that these "freebies" get you hooked with a view to making you upgrade and pay at some point in the future.

Re:Buisness plan

[Stevyn]

It's more like:

Boss: Jimmy, We need a database. What do you recommend?
IT Guy: Well, I used Oracle for a few projects in college. I'm comfortable with it.
Boss: Good! Let's get it. Done. Who wants take out?

Oracle on cheap web-hosters?

[walterbyrd]

To me, it seems like one of mysql's greatest strengths is that it will run on cheap web-hosts. So a lot of F/OSS web apps, like CMS systems, run on mysql.

I don't see the situation changing because these semi-free testing version databases.

They will have to...

[squoozer]

...lever Postgres from my cold dead fingers before I give it up. Well, that or pay me to switch :o).

I don't doubt for one minute this free version of Oracle will be just as hard to maintain and use as the non-free version so why would anyone developing a small application use it over Postgres or MySQL?

Why MySQL?

[ChrisF79]

For me it's pretty clear. MySQL is very easy to set up and use, and I'm already comfortable with it. It really does everything I need it to (plus a ton of other things I'm sure are there that I just don't use). So for me to switch to Oracle really serves no benefit. The reason I bring it up is because I'm sure there are a lot of other stubborn people out there like me that won't be willing to switch or have no motivation to do so. I think MySQL will be just fine, regardless of the Oracle opportunity out there.

Truly a little brother?

[aftermath09]

Can XE handle a backup/restore from a full production enterprise edition version of 10g as long as it falls within usage restrictions (eg. max size 4Gb)? this would be quite handy, as I hate the resource requirements of 10g.

Re:Truly a little brother?

[aftermath09]

Sorry for replying to my own comment. pushed submit too fast. according to the XE faq: Any application developed for Oracle Database XE will run completely unchanged with Oracle Database 10g Standard Edition One, Standard Edition, or Enterprise Edition; so u can backup a DB created in XE, and restore it on the bigger ones. I still wonder if you can do it the other way around. the importance being that u could then do development on something like a laptop without needing to use another small footprint db like msde, mysql or postgresql.

Not enough swap space???

[mikaelhg]

The amount of swap space on a default RHEL4 / FC4 / CentOS 4 with 2GB of memory seems not to be enough for Oracle 10g XE.

"This system does not meet the minimum requirements for swap space. Based on the amount of physical memory available on the system, Oracle Database 10g Express Edition requires 3039.0 MB of swap space. This system has 2046 MB of swap space. Configure more swap space on the system and retry the installation."

A simple workaround.

Re:Buisness plan

[grazzy]

Isn't it more like the above and this:
IT Consultant (two weeks later): Hey guys, why are we using a $10.000,00 solution for storing 100 rows? There's two perfectly working opensource solutions for free.. While you guys to fetch the pizzas I'll merge our database..

Funny posts on this article

[Anonymous Coward]

It is quite funny read posts in this topics comparing Oracle to MySql or Postgress. Anyone that worked in a real database (read: not a crappy forum in a webpage) knows the difference, MySql and Postgress are nothing but toy databases, Oracle is the real thing.

Some reasons for using Oracle

[jandersen]

Size and scalability aren't the only reasons for using Oracle. Oracle has a lot of features that you don't find in any free database, as far as I know. Oracle's SQL, for example, is VERY powerful. You may think that a 'select' is just a 'select', but try to compare them. There are many other features in Oracle that can come in handy even for small databases, like database links and the several kinds of indexing etc.

I have been on the lookout for a free-ish RDBMS that I could develop against for a long time, but I always come up against limitations that I can't live with and that I don't have to live with Oracle. I say, this is fantastic news.

It's not the death of mySQL

[digitalgimpus]

It can only run servers with one processor, with 4GB of disk memory and 1GB of memory.

This means for a true free system, you still need a DB with no limits.

What this will allow is for apps written for just Oracle to run a bit for no cost.

This doesn't really change things.

XDB

[ducttapekz]

Will this version containt XDB? So far Oracle has had the only XML database worth two poops.

Re:XDB

[ducttapekz]

I didn't get an answer so I downloaded it and it has all of the xml database support. Very cool. There are no other databases out there that have such great XML support and performance.

Re:What does this mean for [...] MySQL and Postgre

[Decker-Mage]

I agree completely and have pointed this out elsewhere when the subject has come up. This product is competitively targeted at MS and IBM. I don't think Oracle even thought two seconds about MySQL, postgreSQL, or BDB (which is sad as they are competitive in my book for certain market segments). I'll need to examine it more closely but from first glance I'd say SQL Server 2005 Express Edition is the feature winner here but I'll have to kick the tires, throw some tools at it, and see. It's certainly a huge monster, that's for sure.

Amen!

[Aldric]

My MySQL system ticks along just fine. They could offer the whole thing for free and I wouldn't be interested - I don't fix what ain't broken!

Re:administrative nightmare

[nexus987]

Oracle is a complete nightmare to administer. They have a three day (I believe) class just to learn how to back the friggin thing up! Compare this to Sybase and MSSQL, where you just do a "dump database..." "load database...". Sure, Oracle has a pile of esoteric features, but most people only use the basics.

The limits are not friendly

[pcause]

The article says that the limits on use are 1 CPU, 1 Gb of memory and a 4 Gb database. In a year it will be hard to even find a single CPU system becuase of the move to dual cores. And a 4 Gb database isn't much if you are using it to store media elements and not just text or data.

Seems that Oracle wants the press release but doesn't want to deliver a real, usable product. And, btw, previous verisons of this engine have not been 100% compatible with Oracle's main commercial database.

Try MySQL, Postgress or even MS's free version of SQL Server

What does this mean for the future of MySQL and Po

[segedunum]

Nothing. This is grossly crippled, just as MSDE is. It's been funny to hear all the Microsoft fanboys hype MSDE as being far better than MySQL, yada, yada. I assume Oracle will be doing the same with this.

Oracle is already free ....

[CedricVonck]

Like the title of my reply already said ... you can almost download any tool of Oracle for free. (except the Apps Suite) Only for development purposes like stated on http://otn.oracle.com./ You can download the _entire_ Oracle 10g2 Production Relase DB, the developer suite, and so on & .... It is only a problemn when you use it in your company. My 0.02$ Regards

Just more pressure from the open-source industry .

[Sonic McTails]

With Ingress being open-sourced and what-not, Oracle got to be feeling the heat. While it's the only one that can scale to Forture 100 needs at the moment, it's quickly loosing it's gap as PostgreSQL, Ingres, and Firebird all race to get that top position. Now if they open-source Oracle, I think they might have a chance in hell.

swatting flies

[irote]

I know that this is like swatting a fly with a nuclear weapon

That's a game scheduled for release at Christmas.

Now there's a good question... Could you swat a fly with a nuclear weapon? From what I remember, the major destructive force of a nuclear explosion comes from i) the initial passing wave of heat - which lasts a very short time, followed by ii) a huge pressure wave...

it seems to me that a fly should be immune to this. the heat wave is over in instants, and a fly doesn't have many cavities that will explode or implode under changing pressure. and being flicked a couple of kilometres on a pressure wave in the open air shouldn't matter much to a being that small, and especially not to one that can fly in any case.

so it appears that, unless the fly was right at the epicentre and was incinerated, the fly should just buzz on irritating the survivors regardless.

what's slasdot's view on this important matter?

Re:swatting flies

[808140]

Two things: one, flies (like all insects) have a tracheal circulatory system. This means that they have no blood, but instead have a lot of small tubes (trachea) instead of veins and arteries that carry oxygen directly to the cells. As a result, your statement that a fly doesn't have many cavities is, well, wrong. Of course you could make the argument that these tubes are not closed, and so should be immune to a pressure change, but I very much doubt that the air circulating inside the fly's body would react to the pressure change as quickly as the pressure outside being acted on by the nuclear weapon would.

Two, "unless the fly was right at the epicenter and was incinerated" should be "unless the fly were right at the epicenter and were incinerated". I know, I know, grammar nazis annoy everyone, and my grammar is far from perfect, but I just want to point out that the subjunctive mood in english is ... ahem... not yet dead!

(Incidentally, the reason you say were and not was is because your statement is a hypothesis, and is not in fact the case.)

Re:swatting flies

[Anonymous Coward]

From Parent:-

" but I just want to point out that the subjunctive mood in english is ... ahem... not yet dead! "

Are you sure - did you cross check this fact with Netcraft?

Re:swatting flies

[Anonymous Coward]

If flies have no blood, what is the gooey stuff inside of them that tastes so wonderful?

I.T. departments will approve

[mnemotronic]

Many companies have dedicated I.T. departments with at least a few people who are Oracle admins. Once a company has faith in a product (or has built s significant infrastructure upon a product), it becomes very difficult to switch the foundation. Giving away a small edition that would, from the sales pitch, appear to compete head-on with MySql and PostGRE, is a smart move. Take the case of a company where there are developers considering or starting their own "stealth" db apps - I don't mean the Oracle db admins, I mean the engineer who read about PHP and MySql and wants to try a small task-tracking system for his group. She or he knows that the company uses, and has blessed Oracle, so that might lend some weight to using the XE package rather than the "unproven" (from the IT department's point of view) FOSS database. In addition, Oracle has an enormous amount of administrative tools and documentation. MySQL and PostGRE really can't compare in that arena (IMHO). The downside, in my experience, is that Oracle administration really does require someone who has take a heap of Oracle admin classes.

Honey May Be Sweet

[Ranger]

It's good to know that Oracle is releasing a free version that is as user friendly as it's CEO Larry Ellison, that has a profile as slim as a supertanker, and that it has a learning curve as gentle as the St. Louis arch. I'm reminded of an old Gaelic proverb:

Ge milis a' mhil, cò dh'imlicheadh o bhàrr dri i?

"Honey may be sweet, but who would lick it from the top of a briar?"

anyone tried it ?

[Anonymous Coward]

Can you install and manage it without a GUI ?

What features do you need?

[Futurepower(R)]

"... I always come up against limitations that I can't live with..."

May I ask, what features do you need that aren't in SQLite or PostgreSQL?

Another question: I wonder if the free version of Oracle will work with Compiere ERP + CRM, at least for testing?

Here is a Comparison of Oracle, MySQL and PostgreSQL DBMS.

ZDNet article: Oracle to offer free database.

I was not able to find the list of limitations on the Oracle web site. Anyone?

Re:What features do you need?

[Jaseoldboss]

It's in the Data Sheet (pdf warning).

http://www.oracle.com/technology/products/database /xe/pdf/dbxe_datasheet.pdf

• Available on 32-bit Linux and Windows
• Installs using native installers
• English (single byte character set) and International (Unicode) versions available with support for 10 major languages
• Supports up to 4GB of user data
• Utilizes a single CPU on multi-processor or multi-core machines
• Can use up to 1GB RAM (single instance only)
• Fully upgradeable to other Oracle Database 10g editions
• Oracle Text for efficient text-based searches

Re:What features do you need?

[ahodgson]

Wow 4 whole GB of data. I have more than that in PostgreSQL on my workstation at work, never mind the hundreds of GB in our production db.

Re:What features do you need?

[HvitRavn]

"[..] May I ask, what features do you need that aren't in SQLite [..]" Proper administration tools? Being the industry standard for half a man-age? Real professional support? I know, these aren't features. Then again, your question was stupid, so this answer serves you just right :)

Re:What features do you need?

[chris_mahan]

I'm gonna echo that.

4 GB? You can't even store 1 measly DVD.

They should call it Oracle 10g Teaser Edition.

For a simple web site for a specialty shop, Oracle ould be an overkill. MS Access would be better for that.

For a dedicated web services, well, who knows how large it would grow to? Can't be held back by an arbitrary limit in the software, sorry.

Re:What features do you need?

[Anonymous Coward]

No kidding. Pretty sad that Oracle wants to push this out there when it barely competes against MS Access, let alone MySQL or postgreSQL. What it does it gets people used to using the Oracle suite of RDBMS mgmt tools, rather than the alternatives. It gets them thinking that Oracle is pretty cool for just giving this away. But it's essentially trial-ware. For any serious usage it's completely inadequate. The one thing I can see as a positive about this is the fact that Oracle has a lot of great extensions on top of ANSI SQL that I've not found on other mid-sized RDBMS applications. I'm thinking of things like their rollup functions, for starters. Very handy when working with data in a tree hierarchy.

Re:What features do you need?

[Cromac]

It gets them thinking that Oracle is pretty cool for just giving this away. But it's essentially trial-ware. For any serious usage it's completely inadequate.

I'm sure that's exactly what their marketing department is planning on. They still want to sell databases so of course they're not going to offer a free version that will cut into their potential market very much. It's just like the SQL Express version that Microsoft is coming out with, sure it will work for medium to small, light load databases but they still want to sell you SQL 2005. If anything this move by Oracle is an attempt to compete with Microsofts free version more than mySQL.

Re:What features do you need?

[MightyMartian]

They should call it "Oracle Crapshit Version". Oh yeah, I'm gonna toss out MySQL for some hobbled piece of garbage. In the old days we called this demo software.

Re:What features do you need?

[Not The Real Me]

MS Access has a 2 GB database limit. So, it's 1/2 the database that Oracle 10g is.

Re:What features do you need?

[einhverfr]

SQL-Lite and Oracle are not competitiors. SQL-Lite and MySQL are.

Now, PostgreSQL and Oracle.... Well, you have a few features in Oracle that are not in PostgreSQL at the moment, including MERGE and CONNECT BY though the latter has a non-standard workaround acailable, and the former can be emulated where needed using serializable transaction level and delete/insert statements. Both these items are on the TODO list for PostgreSQL btw.

A second major area that Oracle has some advantages is in parallel query execution and data warehousing. Bizgres is working on both these issues (presumably CUBE is also on the roadmap).

However, in general, for 99% of applications, I think that PostgreSQL is a generally better choice than Oracle.

Re:What features do you need?

[chris_mahan]

And 2 GB is plenty for the mom&pop.

Re:What features do you need?

[Anonymous Coward]

"Oh yeah, I'm gonna toss out MySQL for some hobbled piece of garbage." Where do you get that 'hobbled' bit? The size limit ain't imposed by the database ... suggest the brain-mouth/finger action be resync'd, after making the brain work.

Re:What features do you need?

[mgbastard]

• Available on 32-bit Linux and Windows
• Installs using native installers
• Oracle Text for efficient text-based searches

Ah ha - its a direct competitor to MSDE/SQL Express for disconnected access to enterprise information systems. Sync the data to the local oracle db on the traveling users laptop, and then run the transactions later against the live one. Love it. If Oracle provides tools for the syncing without end-user license fees beyond access licenses, they'll have a key competitive advantage. That native installer feature is the giveaway... easy to manage and rollout to hundreds of users...

This won't make me use Oracle for a project ...

[Alain Williams]

Think about it: start a project, oh - it works, great -- oh it is incredibly successful -- better put it on some bigger hardware --- whooops it no longer works because of some licensing restriction !

No: I might choose Oracle because of some features, but this sort of ploy will not tempt me - I don't like surprises when I try to scale.

MS SQL Server 2005 Express

[Futurepower(R)]

Link to SQL Server 2005 Express.

Re:MS SQL Server 2005 Express

[Tassach]

Link to Sybase ASE Express Edition -- voted as best Linux RDBMS by Linuxworld magazine

Link to Sybase SQL Anywhere Developer Edition

Re:MS SQL Server 2005 Express

[ToasterofDOOM]

I know I'll probably get flamed for this, but here goes.
 
My dad uses Oracle at work, and says that it is an outstanding database. He also says, however, that Oracle support staff is a bunch of very arrogant people, overly proud of an already noteworthy product. His company is moving to MS SQL server and they say that it has passed all obstacles with flying colors. He says it is stable, scalable, and an overall good product. MS Can do stuff right and, as of late, they have been doing more and more things right.

What about Firebird DB??

[terryfunk]

Firebird too is as robust if not more so than Oracle. I believe the whole issue, and one that Oracle is counting on is 'mindshare'.

No trial expiries at all if you develop on Oracle

[killmister]

>> It means they can deploy and test against a running version of Oracle with no need to worry about "developer program" memberships, trials that expire, and similar crud. Sorry, but you can do it right now. Developer licence for Oracle is free, just download and run it for developing purposes - no licence expiries at all. Ed

Mutiple benefits

[Sierpinski]

I always enjoy hearing this type of news from a software company. In addition to the prospect of boosting their future sales by getting users, managers, companies, etc. used to their product, it also serves another purpose, both to the benefit of the company and prospective users. It allows people (like myself) who otherwise could not afford such software (nor usually has a reason to purchase the pricey package) to learn the software, so they can make themselves more marketable. Maya did a similar thing with their personal learning edition. The tradeoff for them was that it had a large watermark on every generated frame. The end result? People had basically fully-functioning software to use and learn on for their own purposes, yet prevented (for the most part) any usage in the production or commercial industry. While I still haven't devoted much time to learning Maya, the opportunity exists, and the thought of being able to learn the software without shelling out the thousand+ price that the commerical package costs is a nice one.

MySQL has ease of install and use...

[EMR]

MySQL is a very easy to install, setup, and manage database. Oracle, on the otherhand, is a PITA to install. MySQL's stong points is that it is an easy to use, FAST database. And with the new 5.0 release MySQL finally has bigger set of "enterprise" features. while still maintaining the speed and ease of use of prior versions. Oracle is a database for a different target audience than MySQL. They are begining to provide many of the enterprise features, but for some reason *enterprise* applications seem to like hard to use databases.:-D

But Oracle's SQL Language Sucks

[Prototerm]

I haven't had the chance to play with Oracle 10, but I've found the SQL dialect in the previous versions of Oracle to be just awful. Not completely SQL92 complient at all. Sure, the Object Oriented features are nice, but I'd rather have the same powerful SQL language that databases such as SQL Server and DB2/UDB have (If you've only worked with DB2 on the mainframe, you won't know what I'm talking about. DB2 on the mainframe's not the same thing as UDB). It appears that Oracle wants to go their own way instead of following standards (gee, who does *that* sound like, boys and girls?).

I often have to write code against multiple database engines, and prefer to write the SQL once, and tweak it a little for each DB as needed, usually finding a common syntax that works with most DB's. With Oracle, however, I'm forced to write lowest-common-denominator SQL that's little more than bare "select" statements, and let the application do the heavy lifting. It's annoying to waste a multi-million-dollar Server with 50-60 big fat processors and lots of memory and do the work on a clerk's PC that is so old it can barely run Windows (how about a 450 Mhz Pentium-II with 128MB of Ram running Windows NT4!) The alternitive is to write two apps: one for Oracle, and the other for everyone else.

Re:But Oracle's SQL Language Sucks

[CedricVonck]

Sorry mate, but from 9i1 or 9i2 you can use Ansi 99 // Sql 99 syntax like inner join, left outer join.

From 10g Release 1 you can use regular expressions to search data... ;-)

Re:But Oracle's SQL Language Sucks

[anvil {UK}]

Well 9i which was released like 3 years ago or something introduced more or less ANSI99 compliant SQL.

You also seem to suggest that SQLServer and DB2 are completely ansi compliant, this is just wrong. No vendor is completely ansi compliant.

What is the use of "Oracle XE"?

[lorcha]

The way I see it:

1. MySQL: Simple to install and learn, good enough for many purposes.
2. PostgreSQL: More feature-complete than MySQL, but takes longer to learn and is more difficult to administer. A good, solid, open-source database.
3. Oracle 10g: Enterprise DB. No, it is not just an expensive substitute for MySQL+replication.
4. Oracle XE: ???

Seriously, there are already good, truly-free alternatives to Oracle XE that developers already know and trust. I would not advise a client to use Oracle XE because who knows if it's going to be around in 10 years? Sure, Oracle's flagship database isn't going anywhere, but Oracle could decide that this XE product isn't generating enough revenue to justify its continued development and can it.

That's about the last thing you need. Do have a bunch of imporant data trapped in an unsupported datastore.

Won't replace MySQL or Postgresql: OS Portability

[laffer1]

The problem with Oracle is that it runs on a very small number of operating systems and processor architectures. Open source databases like MySQL run on WIndows, Linux, BSD, OSX, Solaris, etc. MySQL can run on PPC, Sparc, IA32, AMD64, etc. For example, I could not port my current java based website (tomcat, servlets/jsp) from MySQL 4.x to Oracle simply because my target platform (FreeBSD 5.4) is not supported. Trying to get Oracle to run in FreeBSD's linux emulation is impossible. Heck i downloaded Oracle 9i and 10g for linux and tried to get it up on a redhat EL3 workstation box and that was a real chore in itself. Oracle's installer sucks. Its java based and is VERY picky about the java vm version it is running under.

I'm not fan of MySQL at this point, but it is easy to get running and with MySQL 5 out the door I may finally get stored procedures provided php isn't totally broken for my other sites.

Oracle if you're listening, think about fixing that installer and get some more OS support. OS lock-in sucks and if i have to deal with that I might as well use MS SQL Server which I like a great deal over MySQL.

Downside: not every platform

[wandazulu]

The beautiful thing about mysql is that I've run it on windows, os x, linux (ppc, x86 and mips (ps2 running linux)) Solaris x86, and freebsd. Because you can actually get the source and compile it, you have a transactional database that runs on practically anything. This version of oracle is currently available for Windows and Linux. For some of the groups I work with, this basically makes it a non-starter (there really are some mac-only shops out there who will not think to buy a pc just to run the db).

Re: "Oracle is a rubbish dinosaur"

[smparadox]

Ah, THAT explains why my place of employment uses it then!

Gratis not libre software.

[jbn-o]

Yet another proprietor trying to separate you from your software freedom. From the FSF:

"Sooner or later these users will be invited to switch back to proprietary software for some practical advantage. Countless companies seek to offer such temptation, and why would users decline? Only if they have learned to value the freedom free software gives them, for its own sake. It is up to us to spread this idea--and in order to do that, we have to talk about freedom. A certain amount of the ``keep quiet'' approach to business can be useful for the community, but we must have plenty of freedom talk too.

At present, we have plenty of ``keep quiet'', but not enough freedom talk. Most people involved with free software say little about freedom--usually because they seek to be ``more acceptable to business.'' Software distributors especially show this pattern. Some GNU/Linux operating system distributions add proprietary packages to the basic free system, and they invite users to consider this an advantage, rather than a step backwards from freedom."

It's not free, it's gratis [nt]

[Arker]

[nt]

Re:Free?! Not cheap enough.

[briansmith]

Oracle DBA'ing isn't that hard on a database this size, especially with all the automatic configuration that 10g has. The Oracle database (NOT the app server) is one of the few pieces of "big" software that I have seen that gets easier and easier to use as new versions come out.

PHP and Oracle

[sinkemlow]

Exactly. I'm an amature PHP coder, which means I cut my teeth on PHP and MySQL. Last year, however, I was forced into a position where I needed to working with an Oracle database. I spent a couple of hours reading up on OCI8 and after a quick recompile of PHP, I was working with the Oracle database through PHP.

After getting the basic framework of the PHP application together, my reaction was, "Wow, so this is what they mean when they say *real* database." I had very little difficulties working with the new database, and very quickly began to appreciate the extra features Oracle had to offer. After about a week of working with Oracle, I found it quite hard to fall back into the old PHP+MySQL routine (although the mysqli extension has made this easier).

So I would say the average PHP light coder should have no problems transitioning to an Oracle database. I sure didn't.

Re:administrative nightmare

[stanmann]

So how does Sybase work if you want to do a point in time restore to halfway between to backups? I've been looking for the answer to that question for quite a while. What if you need to rollback to just before a certain committed transaction?

dump and load = export / import

[moscow]

Yes, of course - and you can back up oracle with a one line export command too. If you need to be able to build backups that can cope with pretty much anything - and you need to know how to recover as well - then a three day course shows you how.

These days, even with Oracle, you can click-install-forget and you've got a workable database. But to maintain a corporate infrastructure, you need the whole toolkit - and Oracle lets you tinker with just about anything.

I reckon this is a pretty exciting innovation. I develop using my free developer's copy of Oracle 10g downloaded from OTN. If I can take that into production, then think about paying only if the product grows, it will be fabulous.

Innodb ?

[lesv]

With Oracle's purchase of Innodb, (a key component of MySQL). I find that they are in a sense getting a range of products at the low end. I'm surprised that there hasn't been any speculation on how these two events (InnoDB & a free 10g) will play to help Oracle recapture revenue.

Re:Innodb ?

[Anonymous Coward]

Buying a company only means they want control over it. It's hard to tell whether they want it expand or kill it off. When Microsoft bought FOXPRO, they just want to make sure the FOXPRO develop in such a way that it is not competing with MSSQL.

Oracle just needs to open its source now!

[OneSeventeen]

Unfortunately I still won't be migrating over to Oracle. I am currently running an Open Source server, and would like to keep it that way. I am happy for the non-closed-minded-zealots out there who can take advantage of this and utilize the code they've written on the free development version, but I'm too closed minded to use closed-source just yet.

This will hopefully push OS DBs such as PostgreSQL to work even harder at competing with Oracle, but for my applications I haven't needed anything more than PostgreSQL anyway. Now the question is, if users donated half the cost of oracle each time they installed PostgreSQL, how advanced would it be by now? (that's directed to the Oracle fanatics, who don't seem to be as annoying as us open source fanatics)

Either way, more free software is available, and I'm hoping this will at least squash some of the vendor lock-in caused by Microsoft's SQL server, since businesses can move smaller tasks to Oracle, and use whatever tools they like to interface with it. (as opposed to the insane urging from MSFT to use MSFT products to work with MS SQL data... or at least that's the impression I was given the last time we used it)

10g is supposedly better?

[sheldon]

I have no direct experience with installing 10g or admining it. The last time I tried to work with Oracle was 8i. But at a VSLive presentation last February, the VP of Oracle development made it a point to mention that with 10g the admin tools had been greatly improved, and you would now save something like 30% of your time working with Oracle versus older versions.

Apparently they went and hired a bunch of Windows developers to show them how this works, because they also now have some reasonably good plugins to VS.NET for developing on Oracle servers.

It's nice. Their old anti-Windows attitude had been seriously hurting them in the market.

The EULA is so damn funny

[Anonymous Coward]

I am not a citizen, national, or resident of, and am not under control of, the government of Cuba, Iran, Sudan, Iraq, Libya, North Korea, Syria, nor any country to which the United States has prohibited export. I will not download or otherwise export or re-export the Software, directly or indirectly, to the above mentioned countries nor to citizens, nationals or residents of those countries. I am not listed on the United States Department of Treasury lists of Specially Designated Nationals, Specially Designated Terrorists, and Specially Designated Narcotic Traffickers, nor am I listed on the United States Department of Commerce Table of Denial Orders. I will not download or otherwise export or re-export the Software, directly or indirectly, to persons on the above mentioned lists. I will not use the Software for, and will not allow the Software to be used for, any purposes prohibited by United States law, including, without limitation, for the development, design, manufacture or production of nuclear, chemical or biological weapons of mass destruction.

There EULA is one of the funniest ones I've ever read. Oh! my god let me see if I am "Specially Designated Terrorists". I mean who writes all this stuff. It goes against free software foundation. It goes against the freedom of the people living in Iraq, Iran and the other countried mentioned above.

Sad Statement

[Stone316]

I don't know where to start with a statement like this.. If I had mod points I definately knock it down.

The problem with developers like this guy is that they don't like to think that the DBA has an important role. IT is getting too large these days for people to be able to be good at everything. Its simply impossible. There isn't enough time in a day for one person to architect, design and support most software applications end to end.

Administering an Oracle DB isn't that difficult but its a full time job. Just like development. At first I was a developer but moved to a DBA role... It would be very difficult for me to go back because so much has changed.

This guy either doesn't have the time or smarts to adminster any database environment. Just because Mysql or PostgreSQL are easier to install (not really) and you can 'forget' about them doesn't mean thats a good thing. I'm sure if we heard some of his reasons why he'd never use Oracle it would probably speak mountains.

(For the record I use mysql, postgresql at home and Oracle, SQLServer at work. While I favor Oracle, i'm interested in database technology in general.)

More FUD

[Stone316]

We don't run any clusters here but I wouldn't think of puting some of our bigger environments such as warehouses, financials on anything other than Oracle. As for footprint, I have many oracle databases running under 200MB of ram and could probably get it even lower. It all depends on the application and activity. The first question you need to ask yourself is: Why is mysql so much faster? The second question is: If you need to ask yourself question 1 then you should be talking to someone who knows. Here's the problem with database shops... As a DBA I would love to use the database that fit the job, much how like developers choose the right programming language for the job. ie, your not going to use Java if you need to build a simple script to ftp some files. I'd recommend mysql for very simple applications or websites. For more complex applications but small-mid size i'd recomend postgreSQL or SQLServer. For enterprise class applications, warehouses, where performance is key i'd recommend Oracle. In reality this normally doesn't happen. Why? Its too hard to keep people trained on all these different database platforms. Its even harder to find people that would put in the time/energy to maintain these skillsets.

Re:More FUD

[Anonymous Coward]

I agree that most companies do not want to use more then 1 or 2 databases. I disagree with you and say that a shop should choose only one if possible. Why? For the reason you pointed out - it is too hard to maintain the skills. I spend 90% in Oracle. 10% in SQL Server and have used MySQL, PostgreSQL, and Firebird. I am pushing my company to use only 1 because I cannot be an expert in more then one. I spend an average of 6 or 7 hours a week on keeping up with Oracle realated features, learning and expanding what I already have learned, etc... Those are hours of personal time. How am I suppose to be an expert in another DB?

I can't tell you how many hours I spent wasting time trying to spead up a query in SQL server knowing exactly how to do it in Oracle. Learning the hard way that a inline view in SQL Server acts very different then Oracle's. Trying to explain to the SQL Server developers that the technique they are using in SQL Server (yea - the "create table temp_...) that they use in their SP in SQL Server is exactly what you do not want to do in Oracle. The SQL is very similair in many ways - the underlying technology is not. And learning each DB's takes a LOT of time.

I can build a system in most DB's out there. But I only truly understand Oracle's framework. A side note, I like Oracle because they embrace and work with a number of committies on a open framework. It runs on almost any OS (although not supported on many ike FreeBSD). You can now use Java, .Net, etc... to code your stored procedures. I also like their documentation - all 25,000 pages of it (I read somewhere it is up to 45,000 pages). If you need to know how do to something, it's in there.

I would expand even more - but I have some reading to do ;)

This helps me out!!

[gnu-sucks]

Lots of jobs in the area want people experienced in Oracle. And if I can try it out for free on my own network, then I'm all set!

Thanks Oracle

Of course, that doesn't mean I can't enjoy MySQL for anything worth shaking a stick at...

It's mostly customer perception?

[smchris]

My introduction to C/S databases was Oracle -- stepping directly from Borland desktop dBASE and Paradox. That said, I've been pretty amazed at what PostgreSQL can do -- if you don't absolutely, positively need moment of failure recovery, it seems like PostgreSQL would work great for organizations below that top level of cream that pulls a heavy input transaction load.

But a small-organization free Oracle is somthing I will almost certainly feel I have to look at. It's one of the 800-lb gorilla twins with DB2, right?

It's not MySQL leading the pack

[jgardn]

It's not MySQL that is causing Oracle to fear. PostgreSQL recently announce 8.1. Take a look at the feature set. Compare with Oracle's feature set. The only thing Oracle has on PostgreSQL now are apps. Even then, most of the apps Oracle has are easy enough to code up that it's hard to convince a tech shop to use Oracle over making their own.

Databases are now a commodity.

Oracle?

[jgardn]

Oracle is continuously leads the pack in benchmarks...

I thought the EULA prevented anyone from benchmarking Oracle? That's why no one can publish side-by-side comparisons of Oracle and PostgreSQL. Go ahead and search the internet for a comparison. I can't find any good ones. That's whay I also can't tell you about my own experiences.

All I can say is go try it yourself.

Re:Oracle?

[eric2hill]

Try the TPC benchmarks.

With "Gratis" Free..Software

[WebCowboy]

...you get what you pay for (whereas the opposite is true about "Libre" free software)

I work in the medical field, and a free 4GB-max "validated" Oracle database is a huge boon to a wide variety of medical ISV's.

I can't speak for regulations in the medical industry, but I have had to certify point-of-sale applications that involve communication with credit/debit-card authorisation systems. Trust me--it does not matter if you use Oracle or MSSQL or PostgreSQL as your database engine--you still have to pay the testing fees and do the battery of verification tests for your application before they permit you to connect to the live system.

The only way you'd avoid such an exercise would be if Oracle or others developed and certified the application themselves and you were just implementing it--an example would be if you wrote some mod-perl to call PayPal services. That would not need certification with banks because the transaction is carried out on Paypal's already-certified systems and you are merely calling the function. However, Paypal needed to do much more than use a certified database. Such certification of a database engine doesn't mean the application/solution is automatically good to go. Perhaps there are pre-certified modules in Oracle Financials, but don't count on this "lite" no-cost Oracle data engine being certified for anything of importance.

Validate it, clean up a few features (notably auto-vacuum and passable auto-tuning, maybe some multi-master replication), throw in a simple deployment for ORM or database indirection, and you've effectively moved that commoditization up one layer from the small website developer level.

With respect to PostgreSQL, it seems that it is developing at a relatively quick pace and I don't see how Sun's participation (or lack thereof) will have a notable effect on the development of the advanced features you mention such as multi-master replication, and the tuning abilities of PostgreSQL have improved noticeably in recent releases (and though the manual VACUUM is an annoyance to some, in all but a very few applications it wouldn't kill you to add a line to your crontab file). OTOH, this "validation" of which you speak is not likely to happen very easily. Compliance with industry regulation requires money even when all your developers are volunteers. Furthermore, every release of the product would have to be re-certified.

A "verified" version of PostgreSQL is more likely than mySQL IMHO, but I don't believe you'd ever see such a version freely available--more than likely it would only be available from Red Hat and/or Sun. And, since PostgreSQL is BSD and not GPL any source code changes required to meet regulations need not be made public. Red Hat may be more chartible in that respect, but given Sun's selective commitment to Free software I wouldn't count on it.

Re:With "Gratis" Free..Software

[electroniceric]

I can't speak for regulations in the medical industry, but I have had to certify point-of-sale applications that involve communication with credit/debit-card authorisation systems. Trust me--it does not matter if you use Oracle or MSSQL or PostgreSQL as your database engine--you still have to pay the testing fees and do the battery of verification tests for your application before they permit you to connect to the live system.

To be honest, I'm just starting down that road on the medical side of things, so I defer to your experience. However, I'm pretty sure that you can be asked to validate not just the performance but the design of components of software, like libraries and databases. Which can be harder in a distributed project like Postgres, since there's no clear repository of design documents, nor a particular build control process. Perhaps someone like Pervasive will take to providing the backing for this kind of validation...certainly the kind of money involved precludes a volunteer community from undertaking it. In any event, I'm all for people "certifying" Postgres in whatever way they like - anything that raises its profile is a good step.

Sort of

[einhverfr]

The idea is that you attack MySQL and PostgreSQL by making it easy for developers to use your software for low-end tasks (the sort that MySQL usually dominates) and then offer an easy upgrade path to the expensive version.

Now, this is a bigger threat to MySQL than PostgreSQL. People pick MySQL because it si free (as in beer) and might do an adequate job for (insert light-weight task here). They choose PostgreSQL if they are going to need a Free RDBMS that will offer robust features and grow with them. So MSDE and Oracle 10i Lite are more likely to be a threat to SQL-Lite and MySQL than PostgreSQL.

Re:Sort of

[Anonymous Coward]

Just FYI: "Oracle Lite" is a separate product. The new "XE Edition" of Oracle is based on the same codebase as Standard Edition and Enterprise Edition. Each just has different limits on how you're supposed to use them. Trial and development licenses have been free for all of these, while the new edition allows free production use as well. If the limits are bothersome, SE is $150/user I think.

Oracle DBA is not that hard of a job

[briansmith]

Oracle is easy to admin if:
* you are running Oracle9i or later
* you don't need to use partitioning, clustering, or any of the other $20k options for the Enterprise edition.
* you can dedicate a machine to running it; you aren't trying try to run Apache, Oracle, your mail server, etc. all on the same machine).
* you are comfortable with the operating system that it is running on.
* you install as much RAM as you can so that you can avoid 99% of the performance issues that can be fixed by tuning settings.
* you use automatic extent management (the default)
* you don't use archive log mode
* you can use Oracle's built-in authentication, or your whole network is on Windows and you can use Windows integrated authentication.
* you don't need to run "hot" backups.

The perception that being an Oracle DBA is a difficult job comes from a few sources:

* People that have to manage clusters and/or networks of interconnected databases. These people have a difficult job but I'm sure they will tell you that MySQL, PostgreSQL, and probably even other commercial databases are not an option for their needs.

* People that were DBA's before Oracle9i was released, and follow the practices that they read about in the "Oracle8 DBA Handbook" instead of the best practices documentation that comes with Oracle9i and 10g. Being an Oracle DBA back in the day was more complicated, but now with the versions released in the last three years, if you just do as Oracle recommends you can forget about all the pain that Oracle DBA's use to suffer through.

* People that have no need for a scalable, high-performance database. If you have a database that is so small and simple that you don't need to care about indexing, referential integrity, etc., then documentation for Oracle will look scary to you. But, if your database is not so simple, you will need to do basically the same stuff no matter which RDBMS you are using.

Really, honestly, I bet that anybody that can install Linux and use it successfully can configure and maintain Oracle with an amount of effort that is roughly equivalent to the effort required to configure and maintain any other RDBMS, as long as the functionality that is enabled is the same.

Could be good news for smal office oriented apps

[zeke2.0]

Could having a well known database back end help Open Office or some other contender to compete with 'MS Office for small business' become a viable alternative? Just wondering.

1 socket or 1 core

[Anonymous Coward]

"The latest edition is the same as other databases in Oracle's lineup but is limited in usage. It can only run servers with one processor, with 4GB of disk space and 1GB of memory. Oracle on Friday offered a beta version of the new database for Windows and Linux on its Oracle Technology Network Web site."

It is obviouse they are are referencing a single socket; but does this limit it to a single core? Anyone know?

Use Enterprise Manager

[epgandalf]

If you want a good administration interface, use Enterprise Manager. In fact, Oracle just released a new version: 10gR2.
Press Release

/I work on Enterprise Manager at Oracle.

And then the next lousy programmer comes along....

[zardo]

... where buddy_id like '%38277%' or buddy_id like '%93882%' or buddy_id like '%1838877' or buddy_id like '%88392%' or buddy_id like '%1827332%' or .....

Sybase has done it. It's development only though

[ACORN_USER]

A developer's release will make it easy for people like myself to work on our own environments when our employers have forked out thousands for commercial licenses. The home ground for postgres and MySQL is really the smaller firm/academic institution where you really have to justify spending. Smaller software houses, code-shops and projects will still continue to gravitate towards MySQL and Postgres. The realm where ideology is still cost-effective will not change - and a good thing too.

Thanks for the link, Barry!

[CyricZ]

That's a fantastic link, Barry. Thanks for letting us all know about it.