IBM Sets DB2 Database Free (Beer)

Anonymous Coward writes to tell us that in the hopes of winning more developer interest, IBM has released a free version of their DB2 database. From the article: "DB Express-C is the same database as IBM's commercial offerings but the company places limits on what kind of hardware it can run on. It can be deployed on systems with two processor cores or up to two dual-core chips on Advanced Micro Devices- or Intel-based servers. The memory limit is 4GB but there are no limits on the size of database or number of users. "

Comments

Question for/from the Inept

[eldavojohn]

*puts on his ragged and unused database admin hat*

*paints two large concentric circles on his chest, one inside the other*

*throws off his gloves ... and steps into the circle of death that is known as a Slashdot forum*

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser. All those details aside, why should I switch to either Microsoft's SQL Server 2005, Oracle 10g or DB2?

I'm a huge MYSQL fan ... well, because I've enjoyed their stuff freely for quite some time. Suddenly, everyone (and their dog) is releasing their once-thousand-dollar database in a free and slightly inhibited form.

Why the hell should I even bother thinking about switching to these new databases? And, further more, it seems that most of these newly free databases are not intended for corporate use ... like DB2's memory limit and the castrations of the other databases ... so where do they belong? In the hands of small businesses? A company's "developer camp?"

And could you be so kind as to make the prior assumption that I know I'm an idiot so you don't need to tell me that. Just give me some nice hard facts that are easy to measurably prove one database has an advantage over another.

By the way, thanks ScuttleMonkey for clarifying that it's "free as in beer." Unfortunately it's only noon and now you've awakened my desire to get slammed over lunch.

Re:Question for/from the Inept

[jzeejunk]

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser.

You are underestimating the slashdot crowd. Try posting a link to your website here.

Re:Question for/from the Inept

[1000101]

"so where do they belong? In the hands of small businesses? A company's "developer camp?"

.edu
Let them use it, and they will come.

Re:Question for/from the Inept

[Anonymous Coward]

its much cooler to run Oracle instead of MySQL :-) ...u can proove ur knowledge in enterprise software

Re:Question for/from the Inept

[Saeed al-Sahaf]

I'm glad you're happy with MySQL for your personal web site. But what does that have to do with DB2? IBM is trying to attract developers, not small web site webmasters.

Re:Question for/from the Inept

[GWSuperfan]

I'm not sure I'd bother switching. MySQL 5 supports all sorts of cool features. Combine that with the fact that you already know the product and the decision as to whether or not to switch should be a no-brainer. Unless, of course, you need some feature that isn't in MySQL- but I haven't run into that particular problem yet.

Re:Question for/from the Inept

[AKAImBatman]

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser. All those details aside, why should I switch to either Microsoft's SQL Server 2005, Oracle 10g or DB2?

You shouldn't. In fact, it's mildly surprising you're even bothering with a database.

You should download the big databases if you fit in any of the following categories:

1) You need to develop against or evalute the database.
2) You need to learn the database to increase your marketable skills.
3) You happen to run a high-traffic site and you need features not found in PostgreSQL.
4) You run a site that you expect to grow in traffic and want the option of upgrading to a more powerful version of the software at a later date.

I'm sure that others can find one or two more cases under which these DBs would be useful.

Re:Question for/from the Inept

[OneSmartFellow]

I'm out of mod points, but this should be 'Funny'

Re:Question for/from the Inept

[eldavojohn]

You are underestimating the slashdot crowd. Try posting a link to your website here.

Yes, everyone's intent will be to view my vacation pictures.

Wait a minute ... why is my internet slowing down ... and what's a "port scanner." Well, if Window's firewall is asking me to allow it to come through on port 445, it must be ok.

Re:Question for/from the Inept

[digitaldc]

Unfortunately it's only noon and now you've awakened my desire to get slammed over lunch.

Funny, Paris Hilton told me the exact same thing last week after just 2 bites in to my lunch at Spago.

Re:Question for/from the Inept

[morgan_greywolf]

Two words: Software validation. Database licenses are very, very expensive. You have an app that you've already developed against, say, Oracle or DB2, and you want to to run some test scenarios from a functionality standpoint. You could spend thousands on the software, or you could just keep your testing scenarios within the limits of the free beer version and not have to worry about software licenses. Later, you will probably need to prove out your software from a scalability perspetive and you'll need the real software then, but you don't need to be paying for the license during your year-long release cycle.

Re:Question for/from the Inept

[Belial6]

5) You want to KNOW that your Database will be supported by a global company in the future. Remember, OS2 was only cancelled this year, and then it was licensed to another company for further support. OSS is great, and there is great value in being able to hire programmers to fix and upgrade your applications, but there is also a great deal of value in having a company with IBMs track record totally committed to your data store.

Re:Question for/from the Inept

[drinkypoo]

DB2 is arguably the most scalable RDBMS out there. It was some time ago but when I worked for Tivoli we kicked out some benchmarks and the story was basically that SQL Server (basically Sybase 10 back then) was fastest but least scalable, oracle was slower but much more scalable, and DB2 was slowest but by far most scalable, basically never slowing down any measurable amount no matter how much data you stuffed into it.

Whether this is the current state of affairs, I can't say.

Personally I'm a postgres fan but I'm using mysql for fiddling around with drupal because the mysql support is exceptionally poor. Most all the modules have schema information only for mysql, and sure you can translate it over, but some things won't work quite reliably in all cases. So the answer, as usual, is that you should run the platform supported by your applications.

Re:Question for/from the Inept

[Stone316]

I may be misinformed but I thought if you were developing an application to run on Oracle you didn't need a license? I know thats the way it worked in some of my previous environments.

Re:Question for/from the Inept

[iabervon]

The big databases seem to do better with tons of reads and writes. They also tend to have extensions that let you do things to make certain types of complicated queries a lot faster (such as "for each month, give me the number of rows with a date column in that month and the total of the value column, also the total number of rows, and the total of the value column. But only include rows where the owner column is this value.")

The reason that they're releasing free versions is for developers whose software is used by big companies. If you were working on software for some big company and the database companies didn't offer free versions, you couldn't provide a solution using an expensive database, because you wouldn't be able to test it without paying a lot of money that you won't want to. Then the big company doesn't have a reason to buy the expensive database, because their applications aren't tested with it and probably won't run any faster. When I was working for a company that made a web application intended to be deployed internally by big companies, we used half a dozen free copies of Oracle, and probably wouldn't have used Oracle if these hadn't been available. Of course, these free copies of Oracle were running on workstations and only dealing with example data, not real customer data, so, from Oracle's point of view, the weren't actually doing any database work.

Re:Question for/from the Inept

[Anonymous Coward]

I don't think you need a paid license for software validation from major vendors. Oracle and DB2 were available for free for development and/or internal use for quite a while. I often get IBM newsletters that are offering free DB2 for this kind of work.

It looks like IBM is now giving DB2 free for more than evaluation and development work. And that's what makes it news.

Re:Question for/from the Inept

[GooberToo]

I'm glad you brought up performance versus scalability. PostgreSQL is often considered slow by MySQL fans who fail to understand the concept of scalability. Simply stated, MySQL is typically faster than PostgreSQL with low scalability requirements yet PostgreSQL tends to scale much, much better than does MySQL for both complex queries and highly concurrent, mixed operation loads. Obviously, this is a rule of thumb and not a hard/fast rule. I'm sure there are corner cases (which is often put forward in MySQL benchmarks, which are not reflective of real world applications) in each camp which ignore the rule.

For people that do not understand scalability versus performance, let's put it like this:
MySQL is fast for one user and PostgreSQL, while fast for one user, tends to be slightly slower than MySQL. On the other hand, add a hundred concurrent, mixed operation (aka, not read only) users and MySQL tends to go belly up. At the same time, given the same example, PostgreSQL is happily chugging along; albiet at increased latencies. Of course, this statement is broad and makes many assumptions, but it will hopefully help others understand the concept.

So, given your rankings above, PostgreSQL, tends to find a middle ground between Oracle's performance/scalability ranking. In other words, PostgreSQL tends to scale less than Oracle yet tends to perform better. MySQL, on the other hand, performs fairly fast for read-only databases but scales very poorly.

Re:Question for/from the Inept

[penguin-collective]

Simply stated, MySQL is typically faster than PostgreSQL with low scalability requirements yet PostgreSQL tends to scale much, much better than does MySQL for both complex queries and highly concurrent, mixed operation loads.

Well, those are directions most people do not want to scale into; generally, when people design high performance web sites, they tend to go for simple queries and straightforward loads.

Postgres is a great database because of its features, but for most applications, MySQL is both sufficient and efficient.

Re:Question for/from the Inept

[DrXym]

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser. All those details aside, why should I switch to either Microsoft's SQL Server 2005, Oracle 10g or DB2?

The answer probably is that you shouldn't. MySQL or PostgreSQL are just fine for you. They're probably fine for any shop which just needs a DB for moderate use with no expectation that your DB requirements will grow upwards.

Where the free databases wouldn't be fine is if you're a big outfit who runs hundreds of thousands of transactions and for whom database integrity, performance, scalability etc. are the primary concerns. For them, they'll pay the zillion bucks for a database which has 24/7 onsite support and which runs on every piece of hardware & OS under the sun.

Having said that, then why would anyone use a cut-down "free" version of a commercial DB? Well I can't speak of DB2 / Oracle since I don't use them. But I do use MS SQL Server. I don't particularly like it, but it does more or less what its meant to do. One program I maintain works in networked and standalone mode. In standalone mode it stores stuff in a local MSDE database. In networked mode it talks to SQL Server. Since MSDE is just a cut down SQL Server it means that we can run the same code with either configuration and be reasonably sure it will work. Since MSDE is a free redistributable, we can ship it with the app and the customer can decide which mode to run in. Also, if you're a developer, the cut down version means you can play around with a local copy of the data without screwing up anyone else.

That's the benefit. Though seriously I don't think our app is exactly demanding. Some deployments can have > 1 million rows of data and has some pretty hairy queries, but we don't really need a commercial DB. I feel in my gut that we could move over to PostgreSQL and it would perform as well with that. The reason we don't port is because the risk involved in porting five years of spaghetti to another DB is simply not worth it. I know from experience when it was ported from MS Access the weird and hellish ways that DBs can differ.

Re:Question for/from the Inept

[Anonymous Coward]

ok, here it is 127.0.0.1, have at it!

Re:Question for/from the Inept

[h4ck7h3p14n37]

I'm not sure about DB2, but with regards to Oracle, it's not correct to think of it as just a database. If you are, then you probably shouldn't be running it as there are cheaper alternatives. Oracle bundles a gigantic stack of technologies: JVM, LDAP, geo-spatial, content management, text search, RDBMS, etc. Here's the documentation library for Oracle 9i; it's massive.

Re:Question for/from the Inept

[Lithos]

There are many companies I work with that use MySQL for data warehousing applications that just throw too much data at the database, and suffer through MySQL's poor scaling performance (and horrible subquery performance) because it's the only free solution out there. I see this as being a good alternative, provided the rest of the database's functionality is intact. Most of these companies don't throw more than the listed server spec limits at their database anyway.

Re:Question for/from the Inept

[zimus]

Amazing! Your website looks IDENTICAL to mine!

Re:Question for/from the Inept

[Anonymous Coward]

You're jolly well right that the free versions of commercial databases belong in "developer camps". That's where all the real work ($$$) takes place. I think you're a troll waiting to happen.

Re:Question for/from the Inept

[plague3106]

Because I see employers that want Sql Server or Oracle skills, yet haven't seen any that require MySql?

I bet if you took the time to port your database to one of those two and did some benchmarks, you'd probably see better performance... not helpful if your site isn't busy, but if you want to work for an employer who's site will be busy, proably a good thing.

Re:Question for/from the Inept

[Anonymous Coward]

This is a picture of an http://qnc.noctronium.com/images/trucks.gifOracle/ DB2/Postgres server next to a MySQL server. Honest.

I've done alot of quick and dirty work in MySQL, but once the complexity of a project moves out beyond simple data storage, you'll need to upgrade to a full blown RDMS. Triggers, views, user functions. I know MySQL has been integrating these features of late, but I'd rather use something mature. If I wanted to base my development on beta products, I'd use SQL Server.

Re:Question for/from the Inept

[JavaManJim]

Two items here: trouble with huge queries and is the IBM offer only a tease?

TROUBLE. I had trouble running some of my huge queries on DB2. I had to add SQL to split the queries to return partial results. Took a week to produce a final result set. My queries may not have had sufficient permissions to eat all the CPU cycles.

Then when NCR's Teradata came out my company went to that. Teradat is a whole different kind of RDB. First its a 4way box that indexes the data. Then,like WalMart, we had huge 42 (maybe more) node machines. Teradata ran my queries in two hours. The table was a few hundred million records. While performance is good, the problem with Teradata is maintenance is expensive - maybe 17 million dollars a year.

I think we are being offered a lesser RDB. The best one is IBM's DB2 Universal Database Data Warehouse. Is the jewel that will run more than two processors.

Re:Question for/from the Inept

[fupeg]

If your site is just intended for recreation, then I think the answer is that you should not care if IBM, Oracle, and Microsoft all start offering free databases. You have a free database that suits your needs and you (presumably) already have a good understanding of how to develop with it and administer it.

I work on a startup company on the side. Our application also uses MySQL (4.1.) Right now our database runs on a modest machine (P4 3.4 with 2 GB RAM.) We have had a few instabilities, and they seem to be MySQL related. I've considered switching either to a new version of MySQL, Postgres, or one of those free-ish commercial databases. Everything was written using O/R mapping (Hibernate) so it should be pretty easy to switch from MySQL to whatever. Obviously if our traffic took off and/or these instabilities worsened, switching to something like this free DB2 might be very attractive.

Re:Question for/from the Inept

[Phillup]

u can proove ur knowledge in enterprise software

Even if you can't spell...

Re:Question for/from the Inept

[GooberToo]

when people design high performance web sites, they tend to go for simple queries and straightforward loads.

Who said anything about web sites? Besides, complex is always relative. Frankly, fairly simplistic queries for PostgreSQL can often be considered complex for MySQL, given its optimizer. What does, "straightforward loads", imply? Read only? MySQL's bane is concurrent reader/writers to the same tables.

Since you brought up web sites, believe it or not, most web sites require something other than single user, read-only access to their database, even from a web site. Forums, blogs, and even fairly complex, end-user applications are all common place these days. It's the application that dictates the complexity of any involved queries. Like it or not, MySQL is a poor bottom rung for all but the smallest and most simplistic of database driven web sites.

Re:Question for/from the Inept

[Anonymous Coward]

I agree Postgresql is better scalability wise and runs unrestarted without problems for months!

Since people were asking to see otherpeople websites :), this is a stock photo site that runs Postgresql happely-ever-after...
http://www.goodstockimages.com/

Someone said this public licence is only for developes but I think that is not correct, what developers runs a dual procesor :)...

Re:Question for/from the Inept

[kpharmer]

> So, given your rankings above, PostgreSQL, tends to find a middle ground between Oracle's
performance/scalability ranking.
> In other words, PostgreSQL tends to scale less than Oracle yet tends to perform better.

While I agree with your definitions of scaleability & performance, what would lead you to the above conclusion?

Both postgresql and mysql lack query parallelism, fine memory tuning, any kind of partitioning, etc. Without these features postgresql is *at best* only faster than oracle on index-oriented extremely transactional applications. And I'd say it was pretty safe bet that Oracle would stomp postgresql in that space as well on reasonable hardware (ie, not desktops, etc).

Note, of the two databases I would still rather work with postgresql - it's a fine product and allows me to avoid dealing with oracle personnel. It's just not as fast.

> MySQL, on the other hand, performs fairly fast for read-only databases but scales very poorly.

Scaleability should also consider economics, and mysql can cost. It all depends on how you use it. At what? $500/year/database it can easily be more expensive than a small oracle/db2/etc database (which after the first year will drop to 18% or so maintenance). Given that cost postgresql could pull ahead - even assuming equal performance.

And compared to db2 or oracle, then for the classic read-only applications (reporting) mysql will fall *far* behind. Again, the lack of parallelism & partitioning can easily lead to a 40x difference in performance between db2 & mysql. And while hardware costs are dropping, they aren't dropping so far that you want to spend 40x as much on a server in order to support mysql.

Re:Question for/from the Inept

[Anonymous Coward]

Heh. In the future, try something like:
http://127.35.119.38/

Works most of the time and is more believable.

Re:Question for/from the Inept

[walt-sjc]

It's interesting... We use both mysql and postgres on one of our complex sites, for different things. Both are quite good, but one limitation of postgres we kept running into is getting more than 1000 connections - seems to be some kind of hardcoded limit. Mysql on the other hand happily (with tweaking) handles several thousand connections. It's forced us into some hacks like splitting postgres over multiple DB servers (lack of built-in replication is a total PITA.)

Re:Question for/from the Inept

[GooberToo]

Both postgresql and mysql lack query parallelism, fine memory tuning, any kind of partitioning, etc. Without these features postgresql is *at best* only faster than oracle on index-oriented extremely transactional applications. And I'd say it was pretty safe bet that Oracle would stomp postgresql in that space as well on reasonable hardware (ie, not desktops, etc).

Actually, unofficially of course (thanks to Oracle's license), PostgreSQL is well known for beating Oracle in many small to medium size databases. Feel free to ask around. It's not hard to find people that performed their own shootouts of PostgreSQL vs Oracle and picked PostgreSQL because the performance was equal or better at a fraction of the cost. Having said all that, Oracle does tend to SCALE far better for all the options and features you mention. Simple fact is, parallel queries are not going to save you anything on simplistic queries...etc..etc... Simple fact is, all of the high-end enterprise features that Oracle offers comes with a price. The benefit of paying that price is super scalability. ...And no, you would not be the first Oracle DBA (assuming) to be shocked that PostgreSQL can often keep pace if out outpace Oracle for many smaller (small - large; with very large usually going to Oracle) deployments.

If you are curious, check out the PostgreSQL archives and white papers. You'll find many people that contribute to PostgreSQL by actually picking another DB...and the developers are fine with that. What does that mean? It means there is no shortage of people comparing database x to PostgreSQL and asking the developers how to get better performance and/or sharing that PostgreSQL was picked for their project because it out performed x. How does that help? Well, if people are not picking PostgreSQL, then they want to know how and why. After all, it may help identify areas which need further attention or use cases they had not previously considered. Of course, the flip side is, you will find many people thrilled about PostgreSQL's performance compared to other database...and yes...other sometimes equals Oracle.

Re:Question for/from the Inept

[GooberToo]

AFAIK, PostgreSQL has no such limit. In fact, it's far more likely that you have an OS limit that was causing hte problem. Having said that, depending on the nature of your applications, DB connection pooling should be seriously considered. And FYI, PostgreSQL has replication both freely available and commerically supported now. It has for a while now.

What did the very helpful guys in the PostgreSQL mailing lists have to say about your scaling issue?

Re:Question for/from the Inept

[Xurbax]

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser.

You are underestimating the slashdot crowd. Try posting a link to your website here.

Don't you mean overestimating the slashdot crowd?

Re:Question for/from the Inept

[kpharmer]

> Actually, unofficially of course (thanks to Oracle's license), PostgreSQL is well known for beating Oracle in many small to
> medium size databases. Feel free to ask around. It's not hard to find people that performed their own shootouts of PostgreSQL
> vs Oracle and picked PostgreSQL because the performance was equal or better at a fraction of the cost.

In my experience this usually occurs because it takes more expertise to tune oracle than mysql, postgresql, or sql server. But this isn't to say that Postgresql isn't fine for many applications - the performance that oracle/db2/etc offers isn't always necessary.

> Simple fact is, parallel queries are not going to save you anything on simplistic queries...etc..etc...

Actually, I find that in many applications being developed today it is no longer adequate to merely rehost an app from one technology to another. It has become necessary to deliver some analytical components within even simple applications. For example, in a typical simple transactional application now you may log everything that occurs (ten years ago you never would have done that), and provide totals, running averages, etc to show data broken out over time by category, etc. This isn't rocket science, but requires a great optimizer, query parallelism, and partitioning to run quickly.

> You'll find many people that contribute to PostgreSQL by actually picking another DB...and the developers are fine with that.

yep, postgresql has a great community

Re:Question for/from the Inept

[jbplou]

For a little crappy web site it doesn't matter what you use. However if you play the small free versions and you interview for an entry or mid level job using one of the big db vendor technologies knowledge of one these might help in the interview. Make no mistake you aren't going to become an enterprize architech by using one of these and putting it on your resume. But you'd be surprized, there are positions at places I've worked that didn't get a single applicant who ever used the primary technologies being used(sometimes it was common stuff like ASP or Oracle) put there were still many applicants.

Re:Question for/from the Inept

[spasticus74]

*paints two large concentric circles on his chest, one inside the other*
how else can you have two concentric circles?

Re:Question for/from the Inept

[Jamesday]

Is it really fair to call Wikipedia, Yahoo and Google the smallest and most simplistic of database driven web sites? They are all in the top 20 sites in the world with MySQL as a mission-critical database server behind their database-driven sites. It's a little hard to get any bigger than these web sites.

Wikipedia isn't particularly big, with around 400 million rows in its highest row count table and peakimg at 10-20,000 or so queries per second. That 400 million row table is typically being updated many times a second while also being queried many times a second, all inside transactions in response to live end-user requests.

The current InnoDB row counts for the master Wikipedia server, one of five in the main set, are:

Number of rows inserted 969608627, updated 173250135, deleted 798684629, read 154067949289
25.67 inserts/s, 13.67 updates/s, 8.56 deletes/s, 1600.01 reads/s

The averages are over 96 days, from when we switched this one to being master.

Re:Question for/from the Inept

[Notrace]

About partitioning in PostGreSQL ...
It can be achieved using inheritance. See http://www.bizgres.org/assets/docs/html/tblpartn.h tm#wp135195

PostGreSQL does not provide the "CREATE PARTITION ..." - statement yet but I bet it is only a matter of time.

Notrace

Re:Question for/from the Inept

[jadavis]

Well, what starts out simple can quickly become complicated when you need performance. If you need a summary view of a table that's a few GB, MySQL can't help you much there. Disk I/O is only so fast. What you need is a materialized view, and PostgreSQL allows you to do that (with triggers, &c.).

Not only that, but sometimes you want to insert data one way and look at the data another way. The "straightforward" way is using PostgreSQL's features.

So by omitting features, MySQL often makes things much less "straightforward".

Re:Question for/from the Inept

[jadavis]

Both postgresql and mysql lack query parallelism,

Wrong. Most query work is done in parallel, hardware permitting. Queries cannot be run on multiple CPUs at once (at present), which is what you're referring to, but that's very deceiving. CPUs are generally not the bottleneck, in databases it usually has a lot more to do with RAM and disk.

[PostgreSQL lacks] any kind of partitioning,

Wrong. Have you seen version 8.1 with constraint exclusion? It's been out for a while now.

Without these features postgresql is *at best* only faster than oracle on index-oriented extremely transactional applications.

Where's the evidence? If you put PostgreSQL on nice hardware with RAID and perhaps replication, just like Oracle, I would expect the results to be fairly mixed and inconclusive. Some things Oracle would win, some thing PostgreSQL. I'm sure there are problem domains where PostgreSQL just won't fit and Oracle will, but I don't think you acurately characterized those domains.

Re:Question for/from the Inept

[kpharmer]

>>Both postgresql and mysql lack query parallelism,
> Wrong. Most query work is done in parallel, hardware permitting. Queries cannot be run on multiple CPUs at once (at present),
> which is what you're referring to, but that's very deceiving. CPUs are generally not the bottleneck, in databases it usually has a lot more to do with RAM and disk.

When referring to query parallelism I'm talking about it in the same context that you'd use on db2, oracle, informix, teradata, sql server, and probably sysbase: having the database engine split the query into parts and run them simultaneously to take advantage of multiple CPUs.

And in my experience (on db2, oracle, informix) this leads to linear performance improvements - at least up to 4 CPUs. This is the difference between 60 second tablescans and 15 second tablescans - a huge difference.

> [PostgreSQL lacks] any kind of partitioning,
> Wrong. Have you seen version 8.1 with constraint exclusion? It's been out for a while now.

I've used the same methods with oracle, db2, and sql server. I'd definitely use it if I was using postgresql today, but in my experience every implementation has been a pain in the butt to use.

> Where's the evidence?
don't look at me - there's no TPCs for postgresql

> If you put PostgreSQL on nice hardware with RAID and perhaps replication, just like Oracle, I would expect the results
> to be fairly mixed and inconclusive. Some things Oracle would win, some thing PostgreSQL.

I'm sure you're right - there must be some combinations of hardware and application domain in which postgresql would come out on top. There certainly are combinations of application domain, business size, budget, availability requirements, data volumes, and team skillsets - in which postgresql comes out on top.

> I'm sure there are problem domains where PostgreSQL just won't fit and Oracle will, but I don't think you acurately
> characterized those domains.

The domain I mentioned above (reporting, data warehousing, business intelligence, data mining) is one that I've got fifteen years of experience with. I'm sure I'll eventually be using postgresql for this as well, but think there's little value in fooling ourselves about how honest oracle sales staff is, how fun to work with ibm is, or how well suited to large data postgesql is.

Re:Question for/from the Inept

[GooberToo]

Yahoo and Google are mostly read only databases...they update using a backend process, specifically to avoid many MySQL short comings. I don't know the full details; though I do believe it's at off peak hours. Specifically, Google is heavily distributed so a large number of updates can be done on a single box, at one time, while impacting a small number of potential queries.

Given the stats you quote for Wikipedia...it's pretty obvious it's *mostly* a read only database; which is MySQL's sole sweet spot. Heck, look at the stats your self... 16000 r/s with 49 iud/s. The write activity, over the average second makes up only 3% of the total activity. Talk to me again when you have something other than what is clearly a read-only (97%) database. Try MySQL in an environment where the read:write ratio is something like 2:1 or even 4:1 on a fairly complex schema (queries involve many, common tables) and you'll find, as have been well documented on many benchmarks, MySQL can't even finish. And performance wise, falls to the back of the pack of every other RDMBs in the test. And, let's not forget the many bad habbits it forces and crappy SQL implementation it teaches.

Heck, let's not forget /. its self. It takes a huge pounding and survives being /.'d. But then again, many serious comprimises were made to make it survive and it still periodically falls down. Heck, sometimes even when it doesn't fall down, it grinds to a halt.

Long story short, I'm happy Wikipedia is doing well with MySQL but chances are VERY, VERY high that you would be doing *far* better with just about any other RDBMs; be it PostgreSQL, Firebird, or even the likes of PostgreSQL, Oracle, or MSSQL Server.

Re:Question for/from the Inept

[fbg111]

MySQL is fine for things where data integrity is not an absolute necessity, but where it is, for example the db's that store a multinational corporation's complete financial data, you couldn't trust MySQL. I understand MySQL 5 has made improvements in data integrity, but I don't know to what extent. MySQL developed their database for speed and convenience, which is why webdevs like it, and succeeded in creating a very fast db, but sacrificed correctness and integrity for it.

Re:Question for/from the Inept

[penguin-collective]

If you need a summary view of a table that's a few GB, MySQL can't help you much there. Disk I/O is only so fast. What you need is a materialized view, and PostgreSQL allows you to do that (with triggers, &c.).

You still pay the same amount on disk I/O (actually more), you just spread it out.

So by omitting features, MySQL often makes things much less "straightforward".

Be that as it may, the needs of a lot of busy web sites are well-served by the limited feature set MySQL has, and for that limited feature set, it performs very well.

Re:Question for/from the Inept

[penguin-collective]

Yahoo and Google are mostly read only databases

Yes, like most uses of large databases. In fact, relational databases were designed and intended for that case.

but chances are VERY, VERY high that you would be doing *far* better with just about any other RDBMs; be it PostgreSQL, Firebird, or even the likes of PostgreSQL, Oracle, or MSSQL Server.

Sorry, but from first hand experience, I can tell you that that's wishful thinking. If you can get away with using a simpler database (MySQL, in many cases even SQLite), you are better off using it, both in terms of performance and in terms of maintainability. Even if that requires writing extra code in your app because the database doesn't support some complex query type, you're usually still better off using it.

In principle, there could be a high performance relational store that is designed for both fast updates and fast data analysis, but mainstream relational databases aren't it.

Re:Question for/from the Inept

[jadavis]

You still pay the same amount on disk I/O (actually more), you just spread it out.

Not true at all. Let's count some I/O in an example. Let's say you have a 5 GB table, and a few summary tables which keep up-to-date aggregated information about that table. Over time, every update to the main table has to change the summary tables. So, let's estimate that the updates to the summary tables add up to about 10 GB. Then, the summary tables are read 100 times, and let's say that they add up to 1 MB in size total. So, we have 5 GB + 10 GB + 0.1 GB = 15.1 GB.

Now, let's say the summary table wasn't there. We have 5 GB + 5 GB * 100 = 505 GB.

That's a big difference in total I/O. You could hack something into MySQL, I'm sure, but it would either not be ACID compliant or require application code changes. And it certainly would not be as straightforward as it is just using the obvious features.

Be that as it may,

So, you agree then? All I argued against was the claim that MySQL is more straightforward. I never said that people didn't use it successfully. Sometimes people get an idea in their head along the lines of "I don't need fancy bells and whistles, just something lean and mean". And then when it gets down to it, they end up making the situation much more complicated to make up for the fact that the product is missing features.

Some features complicate, and some features simplify. Triggers, views, rules, user-defined functions, user-defined aggregates, user-defined types, and functions can really simplify things when you need them.

I've seen all kinds of strange things that MySQL programmers do to make up for missing features.

I'm not sure whether this is for my argument or against it, but MySQL clearly thinks those features are important because they're trying to catch up now. And some of the important features are already there (or half there).

I'm confused

[AKAImBatman]

How does this differ from the "Free" version they used to give away? I think I still have the install files sitting around somewhere for the 7.1 UDB install that I got off IBM's website. Did they stop offering a free version for awhile, then restart, or is this licensed a bit differently?

Re:I'm confused

[Cletus the yokel]

That would have been a trial version (90 days) most likely...

Re:I'm confused

[AKAImBatman]

Considering that I ran it for about 3 years without it "expiring", I don't think so. According to IBM's site (the details of which have probably been wiped out by this new version), they were offering DB2 to developers for free. Oracle did the same thing. Only Microsoft had one of those stupid 90 day trial things.

All I can think of was that the free developer versions went away with version 8.0, and haven't been replaced until now.

Re:I'm confused

[slavemowgli]

All I can think of was that the free developer versions went away with version 8.0, and haven't been replaced until now.

There you have it, then. Why ask a question when you already know the answer? :)

Re:I'm confused

[madman101]

Wrong. IBM has had free versions for years. This is nothing new...

Re:I'm confused

[AKAImBatman]

Because I don't know. I'm just guessing. I was hoping that someone who followed DB2 a little closer could fill in the details.

BTW, I looked it up. It used to be called "DB2 Personal Developers Edition". The homepage is here, and seems to include version 8.2. So really, I'm back to square one. How about you? :-)

Re:I'm confused

[Fire Dragon]

How does this differ from the "Free" version they used to give away?

The "Free" version that they used to give away was for development use. You couldn't use that version on your production system. At least thats how I remember it. I was playing around with it and decided not to install it because of that limitation.

With this new version, it can be used on small production enviroments also. This makes it a lot appealing because of the starting costs of development and moving to production.

Re:I'm confused

[Alarion]

Oracle still does the same thing. You can download their database (full enterprise edition) for free, for development use.

However, this does not grant you a license to use the database in a production environment (say, powering your Config Management software).

The new "free" versions grant you the license to use it in production environments.

I imagine IBM's new free version is sort of the same thing (I read the article, but have not yet reviewed the license).

Re:I'm confused

[networkBoy]

Actually it's quite simple. The developer edition is free for developers. Once you want to go to production (even on a small scale) you are liable for the licence. This makes it free to run your dB in a production environment so long as it's on 4 cores (or 2 chips whichever comes first) and 4 gig or less.
-nB

self taught learning?

[way2trivial]

want a job? job requires knowledge of a specific app? Ka-Ching....

Re:self taught learning?

[Anonymous Coward]

Use Windows, then.

Re:self taught learning?

[ePhil_One]

want a job? job requires knowledge of a specific app?

Not just that. The GP says

And, further more, it seems that most of these newly free databases are not intended for corporate use ... like DB2's memory limit and the castrations of the other databases ...

Actually, two modern dual core CPU's and 4GB of RAM is several times more powerful than the DB2 server that powered our company through the Dot Com boom. This is a limit in name only. Of course, if you're running your business off it, you will still want spring for the support contract

BTW, Dev licenses for DB2 have gone for token amounts for years. And I thought a restrict DB2 for Linux has been available for years...

Re:self taught learning?

[PCM2]

Actually, two modern dual core CPU's and 4GB of RAM is several times more powerful than the DB2 server that powered our company through the Dot Com boom.

Good point. Extremely good point.

I'm wondering to what extent IBM and the others are banking on the due diligence process to win them their licenses... in other words, "I can get a DB2 server with two CPUs and 4GB RAM for const NOTHING or I can pay $SOMETHING and get eight CPUs and infinite RAM and a lifetime contract with Global Services." The company is under no requirement to spend absolutely nothing, so not spending anything starts to sound like poor due diligence. You spend the money and you get the latest greatest whiz-bang machine ... which might be great for running the Vista UI but is really more than you need for your Web site.

Re:self taught learning?

[ePhil_One]

$SOMETHING and get eight CPUs and infinite RAM and a lifetime contract with Global Services."

High end databases are usually priced per CPU, we saved enough in licensing costs to pay for an upgrade from a old 6-way box to a new 4-way that was orders of magnitude faster. And its not like you can't upgrade to a 8 way DB2 license when you need more, or migrate to AIX based systems, or even all the way up to mainframe based systems with hardware accelerated DB2 functions when you need more umph.

features

[MORTAR_COMBAT!]

does mysql do this:

http://www-128.ibm.com/developerworks/db2/library/ techarticle/dm-0511singh/

Re:features

[zippthorne]

Your article uses the word, "leverage." Furthermore, it uses it as a verb. Stop weirding language.

Re:features

[XMilkProject]

No, but the free MS SQL has been doing it (with more features) for years.

*Braces for attack*

Re:features

[killjoe]

MSSQL can't take an XML document and then parse it and store it as a parsed structure. It can return queries as XML but even that hasn't been around very long.

Re:features

[zardo]

Here are the major differences between commercial databases I've used, all offer a free "express" edition:

Sybase Cheaper version of Oracle, much of the same underlying code, optimised for speed although maybe not quite as fast as Oracle. Missing some features like GIS datatypes. Annoying limitations like a 30 character name limit and no full outer joins (only the union trick). Recently integrated with the java virtual machine to do java stored procedures, good for java development. Popular with financial transaction companies, probably due to stable codebase (not a lot changes). MS SQL Microsoft buys Sybase codebase, which is decent, and build into it a lot of developer studio tools and integration, no Java. Oracle Much the same as Sybase, grew out of the same early-era databases, with more features/options (like GIS datatypes) and arguably, it's faster too. Also the most expensive. Popular with Java applications, java integration just like Sybase. Same annoying limitations like 30 character limit on names and no full outer joins. DB2 Move forward with new features unlike Oracle and Sybase, noticably different feel than the classic databases. Java integration also. None of the annoying limitations mentioned earlier. MySQL Lots of advanced features added in newer versions, such as GIS datatypes (by programmers for programmers, as they say). Fast, low memory footprint, free. Widespread support. Biggest setbacks are the lousy query optimizer and other fundamental aspects of the core which may come back to bite you, although I hear it has a well organized design that aids in development of new features. PGSQL Great database, been around long time, used to be slow but now is faster with more popularity due to mysql success. Free and all that. Biggest limitation in my limited experience is the lack of any replication features built in.

I know I may have missed some major issues that I'm not aware of, so feel free to add to this.

Re:features

[LurkerXXX]

Yeah, it's only been doing it since 2000... So just a little over 5 years. That's a ding? Hmm, exactly how many years have you been using XML anyhow?

Re:features

[Anonymous Coward]

One of the most fundimental advantages of Oracle over the others is that it offers read consistent transactions without locking. This is a fundamental ACID principle that is not possible with other platforms. Moving from Oracle to SQL Server or Sybase can be a shock to the system when you realise that to get consistent safe reads you have to wait for any update locks on the underlying page to be released!

Re:features

[Forbman]

Sybase & Oracle couldn't be more different, once you get past that they use SQL.

Re:features

[Q-Cat5]

It seems to be some sort of inverted gerund. Adding -ING to a noun to make it a verb, instead of adding -ING to a verb to make it a noun.

Anagerund?

The process of which would be Anagerunding?

Hey, it's got self-reference, and that makes it stylish.

Re:features

[jpetts]

Oracle
        Same annoying limitations like 30 character limit on names
        Damn', yes!!!
        and no full outer joins.
        Which version are you using??? This has been in there for YEARS. You just have to use the ANSI syntax...

Re:features

[Anonymous Coward]

Cheaper version of Oracle, much of the same underlying code

Is that actually true?

Re:features

[david duncan scott]

It's "verbing," as in, "Hey look, I verbed the hell out of that poor noun!"

Re:features

[GooberToo]

So does PostgreSQL...it's called MVCC. Here's a completely random link I found which appears to explain it. http://linuxgazette.net/issue68/mitchell.html

So really you're saying, "One of the most fundimental advantages of Oracle and PostgreSQL over the others..." ;)

Re:features

[GooberToo]

PostgreSQL has had master/slave replication for several years, freely available. More recently, PostgeSQL' replication has gotten yet more sophisticated, feature rich, and capable. Furthermore, commerical clustering and replication solutions have existed for PostgreSQL for ages.

Re:features

[Anonymous Coward]

There is now a product for PostgreSQL replication called p/cluster. Check http://www.continuent.com/index.php?option=com_con tent&task=view&id=212&Itemid=169
An open source version is also available from http://www.continuent.org/

Re:features

[Anonymous Coward]

The feature list for Sybase, Oracle, IBM DB2 and Microsoft SQL Server needs work.

Sybase
"Cheaper version of Oracle, much of the same underlying code" is incorrect. Oracle is not from the same code base as Sybase. Early versions of Microsoft SQL Server were derived from the Sybase code base.
Sybase "recently integrated with the java virtual machine to do java stored procedures" -- recently ?? Sybase, IBM DB2, Informix and Oracle did that in the late '90s:
Logic in the Database
http://www.sqlsummit.com/Articles/LogicInTheDataba se.htm/

MS SQL
.. "no Java". There are JDBC drivers, but no embedded Java VM or classes embedded in the database (see above article).
Supports SQL-2003 XML data type and XQuery.

Oracle
Supports XML data type and XQuery.

DB2
XML data type is in next version.

MySQL
Does not support embedded Java classes, XQuery or XML data type.

Re:features

[Anonymous Coward]

and it's a real joy on long-running queries that just die when they run out of redo space

Re:features

[DavidTC]

You bastard, you verbed nouny!

Re:features

[msobkow]

I believe the same is true of Sybase ASE 12.5 -- left and right outer ANSI join syntax is supported.

I also believe that 30 character limit might have something to do with the idea of "standard". Some databases extend the standard, but that doesn't mean a good database design assumes the database will do so. Besides, why in the world would you need more than 30 characters for a DDL object?

IIRC the standard used to be either 18 or 24 characters, so 30 is actually pretty generous in comparison.

Re:features

[killjoe]

I call bullshit. MS SQL server 2000 can't take an XML and save it as a parsed node tree.

Re:features

[Anonymous Coward]

To bad it IS a word AND can be a verb.

http://dictionary.reference.com/search?q=leverage

Get a job.

Re:features

[zardo]

Newer versions of Sybase have row-level locking, I dunno, I figured it was common amongst the databases, a progression they all went through.

Re:features

[zardo]

They both sprung about at about the same time, from the same "crowd" if you will, the guys over at Berkeley working on Ingres. That's what I meant by that, none of the other databases really go back that far.

Re:features

[zardo]

By full outer join I mean a combination of left and right outer joins, i.e. *=*

Last I checked Oracle didn't support them, I dunno, it is always possible to do a left outer and right outer and union the results, that has always been my approach when working with sybase/oracle without ever even thinking about it, like I didn't ever expect it to change. Standards, like you say. For a while I wanted for things to change, but obviously they never change.

I want to use longer names for some tables in my Ruby on Rails application, machine generated names for the most part (mimicking multi-to-multi joins between DAO's.

Re:features

[jadavis]

PGSQL
        Great database, been around long time, used to be slow but now is faster with more popularity due to mysql success. Free and all that. Biggest limitation in my limited experience is the lack of any replication features built in.


How is not having integrated replication a limitation? Replication is a complex problem, and requires a variety of solutions. Moreover, if you integrate a replication technology, the releases must be in lockstep with the rest of the code.

Because the replication solutions are not released as a part of the main distribution, the replication solutions are very version-independent and allow a lot of flexibility about their application.

The developers don't see this as a limitation, and have no plans to change it. So if it's limiting you, tell everyone how, and maybe it will be fixed.

Re:features

[zardo]

Another aspect I failed to mention is that there is no defacto replication for PGSQL. I investigated last year and it was a half assed attempt at replication, and a bunch of commercial offerings.

All the other databases have some supported solution, including free replication with MySQL (integrated).

Re:features

[Anonymous Coward]

http://www.databasejournal.com/features/mssql/arti cle.php/2204421
http://www.perfectxml.com/articles/xml/openxml.asp
http://www.microsoft.com/technet/technetmag/issues /2005/05/SQLXML/default.aspx

Re:features

[killjoe]

HA HA. You do realize there is a maximum limit to how much data you can pass into a stored proc and it's smaller then most XML documents real enterprises use don't you?

Yea, I tried that one but it sucked ass in the real world.

Re:features

[msobkow]

I'm a little confused. Wouldn't *=* just be an unqualified select matrix? (I forget the math/set theory term for the construct, but for n rows in one table and m in another, you get a result set of n * m elements.)

Free (Beer)

[Anonymous Coward]

Forget the software. Where do I get the free beer?

Re:Free (Beer)

[femto113]

Here, apparently.

Re:Free (Beer)

[Anonymous Coward]

No, from now on all slashdot articles will have "(Beer)" appended to the title. Hold on tight.

Who cares?

[Anonymous Coward]

I guess if you're waiting to be downsized from some mega corp and want to play around at home with the same crap you use at work this is good otherwise what's the point?

developers,developers,developers

[free space]

The thing is, most developers use the dev tool/database they 'play with'.
So all companies are releasing a "playful developer edition" of their tools, so that developers learn their stuff, play with it,and when they do a serious project, they will ask their boss to buy tool X because it's what they know.

To directly answer your question: you don't have a reason to replace MySql on your site where you post mexico photos. But IBM wants other people to use DB/2 on their personal sites/apps so that they get the habit of using it elsewhere.

Re:developers,developers,developers

[Bloke down the pub]

they will ask their boss to buy tool X because it's what they know.

And the boss buys something else, because mauve has the most RAM.

Downside

[Nerdfest]

The downside of all this is that if you take them up on it you'll be running DB2 ...

Scalable Software

[RevMike]

Why the hell should I even bother thinking about switching to these new databases? And, further more, it seems that most of these newly free databases are not intended for corporate use ... like DB2's memory limit and the castrations of the other databases ... so where do they belong? In the hands of small businesses? A company's "developer camp?"

First of all, the commercial database offerings are far more feature rich than MySQL, though MySQL is getting better all the time. MySQL is fabulous for the hobbyist and small business crowd, but won't be found as a core technology of a major investment bank or backing an SAP installation at a large manufacturer anytime soon.

So the benefit of these free commercial products is that developers can build systems based on these technologies, then scale them easily to both small businesses and large enterprises. No one will start running their self-hosted blog on DB2, but it is easy to build a software product which may target businesses of several different sizes, and using a consistent database engine for small business and large enterprises is useful.

Re:Scalable Software

[charlesnw]

You cite a couple very narrow scenarios in your post. Major investment banks and SAP installations. As far as SAP they released MaxDB which is what they base SAP on. MaxDB and MySQL are very closely related. As far as investment banks I don't know enough about what they use. And they probably wouldn't reveal that anyway. However it is used by a lot of high traffic sites as the MySQL website clearly documents: http://www.mysql.com/why-mysql/case-studies/ The organization I work for runs MSSQL Oracle and MySQL. Guess what database is used as the core of our business and would cause the most impact if it was unavilable? MySQL. Guess which database has given us the most problems in the terms of deployment/configuration/crashes? Oracle. Guess which database hasn't had any issue (at least that I know of) MSSQL. Kind of opposite of what you would expect huh? :)

Re:Scalable Software

[kpharmer]

> As far as SAP they released MaxDB which is what they base SAP on. MaxDB and MySQL are very closely related.

in what way? If I'm not mistaken MaxDB is the renamed SAPDB which is the renamed Adabas - an old crufy pre-relational database from the early 80s. I remember thinking it was obnoxious around 1985.

> Guess which database hasn't had any issue (at least that I know of) MSSQL. Kind of opposite of what you would expect huh?

Not really: mysql's normal behavior is to silently truncate, silently convert, silently ignore problems. It will often happily chug along through invalid calculations that other servers would stop immediately at.

And it is far easier to administer. Oracle used to be a real bear to design and administer for, though that's changed substantially over time - especially with 10g. In the old days if you didn't have the right skills available you could get in very bad trouble with oracle. It's still true, but to a much lesser extent. Part of the reason is the extreme configurability: you can tweak almost everything. With mysql you can tweak very little in comparison - and often memory pools that should be joined are actually separated (wasting valuable memory). But it's simple.

But I wouldn't put my critical business apps on mysql - even if it is simpler to admin. The designers showed a careless disregard for data quality. They've cleaned it up some, but not completely.

Toes, now feet in the water

[fak3r]

IBM has certainly come a long way in the past few years in regards to recognizing the value of open source software. Now that they have a 'real' free offering(they used to have a bastardized one, and some will argue that this is only free as in beer, that this isn't free either, but...) this should change the dynamic of their marketshare. I know most goverment projects now pretty much want DB2 (at least the ones I've been involved in) but this will open the door for all sorts of options, ppl running CMSs, ppl using Databases for virtual mail users...wow, this is actually a pretty big deal, I think.

Now, IBM, when will you offer either a 'naked' Thinkpad, or one that comes preinstalled with Linux (or FreeBSD) for us who want a real workstation?

Re:Toes, now feet in the water

[bigman2003]

Ummm...I guess you don't keep up on tech news much.

IBM no longer produces the Thinkpad, Lenovo does.

Re:Toes, now feet in the water

[fak3r]

Eeek! Yeah, I knew, that, damn. I suppose that doesn't bode well for a LinuxPad from IBM as I'm sure they will "Recommend" XP. Oh well, I think the MacBook Pro, or just the 'non-pro' MacBook, would be nice with OS X and Linux installed. Now don't go and tell me that Lenovo produces the MacBooks... ;)

Re:Toes, now feet in the water

[Minwee]

And yet, for some reason, the brightly coloured letters "I", "B" and "M" still appear on the bottom right corner of each Thinkpad that Lenovo makes.

Re:Toes, now feet in the water

[Kadin2048]

It changes less than you'd think, though. The ThinkPads were made in China before IBM sold that division to Lenovo, and they're still being made in the same factories, using the same tooling, probably by the same people, now that they're Lenovo. From a consumer standpoint, nothing changed.

I always thought it was odd that IBM never offered a Thinkpad without Windows, but I guess they have the same problem that every other manufacturer did -- start offering an alternative to Windows, and watch your 'most favored reseller' license and pricing to Windows disappear.

That said, you can sometimes get used Thinkpads that are only a year or two old on eBay. They're remarkably well-built, and you don't have to pay the MS tax directly. Whether you're still paying it indirectly, when you're buying a machine with a wiped hard drive but which originally sold with Windows, is an open question though.

Re:Toes, now feet in the water

[Alan Shutko]

They did offer a Linux thinkpad for a while. The TP A20 was offered... ISTR it came with Caldera. I don't think it sold well... it had lower specs than the A20p, which is why I got the p and installed Linux myself.

Re:Toes, now feet in the water

[zlogic]

IBM no longer makes Thinkpads. Their whole PC division was sold to Chinese manufacturer Lenovo.

OK - I'll bite - the 2 proc limit is significant

[xxxJonBoyxxx]

OK - I'll bite - the 2 proc limit is significant. Most of the other "here's a free, hobbled copy of the huge thing we hope you will buy someday" seem to have a 2 proc limit. On the other hand, the truly free databases have so many advantages (for me, it's small footprint) over these that they are not worth looking at - I'd never want the FULL version of these databases, even if those, too, were free.

A little too late?

[beacher]

Just some thoughts and possible misconceptions - I haven't used DB2 at all, but I am a heavy Oracle and occasional mysql DBA/user. I just went to IBM's DB2 page and was not very shocked to see their migration page

"Migrate Now! for DB2 Universal Database (UDB) facilitates the migration from Oracle, Sybase, Microsoft SQL server, and additional database platforms to DB2 UDB at a special price. Migrate Now! is an end-to-end offering that includes migration tool kits, no-charge online education, sales teams and resources to assist you in planning and implementing your migration based on IBM's proven methodology."

I think it falls directly in step with IBM's shift in strategy - lower the software cost and generate service based revenues. I don't think I'll be moving my stuff over anytime soon. Oracle on the data warehouse (the app was built before mysql could do cross table updates), mysql on the select only local repository.
IBM may be too late for the vast majority of developers. The ones that offered their products to develop and learn on are the ones that will find some sort of loyalty.

Re:A little too late?

[AnotherDaveB]

I think it falls directly in step with IBM's shift in strategy - lower the software cost and generate service based revenues.

I'm sure you're right. All hardware / service companies want to get a bigger share of the available IT budget, money not spent on software is available for the New Toy.

IBM may be too late for the vast majority of developers. The ones that offered their products to develop and learn on are the ones that will find some sort of loyalty.

If DB2 is the superior product, I'm sure the vast majority of developers will be happy to consider it.

Re:A little too late?

[tolkienfan]

You're mad.
DB2 was the first relational database (although it didn't have a name back then).
It was also the genesis of SQL.

So I fail to see how they can be "to late" in any way shape or form.

The DB2 database has been free (beer) for years for developers only.

This marks a step in that developers are now allowed to redistribute the database.

Do you actually know anything about enterprise scale relational databases, or are you just making this stuff up?

Re:A little too late?

[Anonymous Coward]

I think it falls directly in step with IBM's shift in strategy - lower the software cost and generate service based revenues.

For fiscal year 2004, IBM's revenue was $96.2 billion. Hardware revenue was $31.1 billion (33.2% of annual revenue). Software revenue was $15 billion (15.5%).
Global Services revenue was $46.2 billion, 44% of total revenue.

So ... why would I move from PostgreSQL to DB2?

[Anonymous Coward]

Besides the obvious "cost" reasons, what would DB2 give me that PostgreSQL doesn't?

I don't care about point'n'drool interfaces or other such frippery. I do care about multimaster replication and rock-solid stability.

Re:So ... why would I move from PostgreSQL to DB2?

[AKAImBatman]

Besides the obvious "cost" reasons, what would DB2 give me that PostgreSQL doesn't?

Plenty of headaches? Bragging rights? Optional corporate (un)support? Good karma? Ability to "fight the man" by "being the man"? (You 'da man?)

See? Lots of reasons! :-P

Re:So ... why would I move from PostgreSQL to DB2?

[DARKFORCE123]

DB2 main attractiveness for being free is its feature set. PostgreSQL already has a well developed feature set for a 'free' offering so you may not get as much from it unlike if you migrated from MySQL. No offense to MySQL, but for now its claim to greatness is being 'free' and speed, not having a fully developed feature set.

Bottom line is that this free DB2 should be considered an evaulation kit for a larger deployment more than anything else since its way more than you need for a small system.

Re:So ... why would I move from PostgreSQL to DB2?

[Kenneth Stephen]

Dont know about other PostgreSQL features, but the two that you've mentioned are covered by DB2 with great panache. Furthermore, DB2 offers features like federation of multiple types of datasources (XML, other databases, csv files, Domino servers, etc). It has OLAP capabilities, native XML support, and has proven and published TPC benchmarks available. On multiple operating systems and different high-end hardware platforms. Why wouldnt you choose DB2 over PostgreSQL?

Yes, if you want the source to DB2, you arent going to get it. But unless you really want to be in the business of maintaining / fixing problems in a complex database product, why do you care about that?

'DB Express-C' available on multiple platforms

[sczimme]

A quick search of IBM's site reveals the links to download DB Express-C. (Registration is required.)

Since no one /. ever, ever, ever runs Windows, here is the link to the X86 Linux, 2.6 kernel version. :-)

Re:'DB Express-C' available on multiple platforms

[laffer1]

Multiple platforms? I count 2 without looking at hardware. Linux and Windows... Thats not exactly ground breaking. IBM has to support windows and likes linux. Maybe they should consider other platforms like say Solaris, MacOS X, FreeBSD, something HP makes, etc. With open solaris and the rumors about licensing floating around IBM may wish they had solaris support. Personally, I have a FreeBSD and Windows setup at the moment so it does me little good. Even if i get it to run under linux emulation, its going to be slow. I guess I'll stick to mysql.

Re:'DB Express-C' available on multiple platforms

[mzungu]

Apparently not anymore..

Funny message for a piece of DB software:

2006-01-27 15:26:14.905745R download was not found in the database

--

Re:'DB Express-C' available on multiple platforms

[fbg111]

Didn't you get the memo? No one on /. (except the QA crowd) runs Linux kernel 2.6 either, we're still all on 2.4, waiting for 2.6 to mature and stabilize, and some of us are still on 2.2, waiting for 2.4 to stabilize...

[ducks]

Re:OK - I'll bite - the 2 proc limit is significan

[mordors9]

At least it wasn't as bad as some of the old "free" offers from some companies. Free download..... downloading...... then install it..... oooops, now you have to buy the key...

Re:OK - I'll bite - the 2 proc limit is significan

[Cletus the yokel]

Well - it's up to *two* dual core processors on x86 (I believe Oracle and SQL Server free offerings are just 2 CPU's or one dual-core), and up to 4 GB address space (SQL Server is just 1 GB), plus unlimited db size.

MOD PARENT REDUNDANT

[everphilski]

concentric circles on his chest, one inside the other

Re:MOD PARENT REDUNDANT

[swillden]

concentric circles on his chest, one inside the other

That's not redundant! It's *entirely* possible for two circles to be concentric, yet neither one inside the other! They just have to have the same radius.

You're one of those people who doesn't see a difference between "positive" and "non-negative", aren't you?

Re:MOD PARENT REDUNDANT

[everphilski]

That's not redundant! It's *entirely* possible for two circles to be concentric, yet neither one inside the other! They just have to have the same radius.

assuming one can apply paint with the painstaking perfection of a machine. Oh, wait...

You're one of those people who doesn't see a difference between "positive" and "non-negative", aren't you?

Depends upon the system (you asked the wrong guy, I'm a simulation programmer...)

Re:MOD PARENT REDUNDANT

[Anonymous Coward]

There is a real and very definable reason that "positive" and "non-negative" are not the same. You just have to remember that there is a number that is both "non-negative" and not "positive." You probably realize this, but who are these people you speak of that don't see the difference and how hard is it to explain it to them?

Re:MOD PARENT REDUNDANT

[Frank T. Lofaro Jr.]

A number?

I assume you mean zero.

How about i and multiples thereof?

Question from the Impatient....

[Anonymous Coward]

OK, DB/2 is fine and all, but how about a free (no 90-day limit) edition of Informix Dynamic Server?

Something I once heard from a contractor: IBM bought Informix in hopes of merging the Informix technology into DB/2... but found that Informix was so far ahead of them that there was no way they could do it without a full rewrite.

Re:Question from the Impatient....

[baldass_newbie]

Alright, I have mod points to burn, so I'll bite...

It's because IBM specializes in killing applications. See Lotus Notes, Ami Pro, OS/2, etc.
Cripes, not that long ago, Oracle was dying to gain Informix market share because Informix was the DB. Sadly, they could open it tomorrow and I doubt anyone would blink.
Good job, IBM. Why don't you do everyone a favor and buy Windows?

Re:Question from the Impatient....

[X]

Something I once heard from a contractor: IBM bought Informix in hopes of merging the Informix technology into DB/2... but found that Informix was so far ahead of them that there was no way they could do it without a full rewrite.

That seems more than a bit unlikely. Informix Dynamic Server just went in a different direction from DB/2, and that did indeed make it difficult to merge the two technologies. That said, prior to the buyout by IBM, the general feeling I got from people who knew better was that IBM pretty much had the lead in RDBMS technology (yes, definitely ahead of Oracle). Informix's DataBlade's were where they were ahead (DB/2 had something similar, but it was definitely trailing Informix).

Re:OK - I'll bite - the 2 proc limit is significan

[xxxJonBoyxxx]

Hmmm...RSA security is the only company I can think of that still does this. Anyone else?

They gave other things away too.

[IAAP]

Visual Age IDE for C++, Java. There was some other things too, IIRC. It was in the late 90's that they had this stuff up there. I can't find them anymore.

Communism!

[Uukrul]

Ballmer: And it had, you know, the characteristics of communism that people love so very, very much about it. That is, it's free.

I'm sure Sam Palmisano has one of this posters in his office.

But I already use

[TheDoctorWho]

MySql and Postgrese, thus no point here except for hands on experience to pad the good ole resume with.

Re:But I already use

[GooberToo]

I'm curious. If you already have PostgreSQL in your hat, why do you bother with MySQL? App Requirements? Ease of use? Most comfortable? Tools? Seriously, what makes you pick one in which situation?

DB limits and removed feature sare a big pain

[duffer_01]

Just to create a blank database, Oracle defaults to 1 GB+. With the inefficiencies of the Oracle db, that does not exactly give me much room to play with. At least IBM is not going to have a limit.

Also, I definitely recommend people look at things that have been taken out of these "free" databases. For example online backup is one common thing that gets removed.

Re:DB limits and removed feature sare a big pain

[jedidiah]

What the H*LL are you talking about? If you use the shiny happy tools to build a database, it will fit itself into whatever resources you have. If that means a puny database and a puny shared memory area then it will create you a suitable database. If you can drive without the training wheels, you could easily create yourself a database and instance that will fit on any current x86 laptop with no problems.

            I built databases on my laptop quite comfortably when it still only had 40G disk and 256M ram.

Re:DB limits and removed feature sare a big pain

[Anonymous Coward]

...you'd still fit easily in 256MB:

[oracle@sandman ~]$ sqlplus / as sysdba

SQL*Plus: Release 10.2.0.1.0 - Production on Tue Jan 31 01:51:00 2006

Copyright (c) 1982, 2005, Oracle. All rights reserved.

Connected to an idle instance.

SQL> startup
ORACLE instance started.

Total System Global Area 113246208 bytes
Fixed Size 1218004 bytes
Variable Size 58722860 bytes
Database Buffers 50331648 bytes
Redo Buffers 2973696 bytes
Database mounted.
Database opened.
SQL> quit
Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production
With the Partitioning and Data Mining options
[oracle@sandman ~]$ uname -a
Linux sandman 2.6.16-rc1-git4 #1 Sat Jan 21 16:07:34 CET 2006 i686 i686 i386 GNU/Linux
[oracle@sandman ~]$

Re:DB limits and removed feature sare a big pain

[Anonymous Coward]

sorry but no backup etc pp removed... all in there...

It's a trial copy

[jd]

I've a copy, myself. Still have the snail mail CD & letter with trial code. My guess is that there's a bug in the timelock. I don't recall seeing the trial version available for a while, so my guess is that it got pulled, but I couldn't swear to that.

Somewhere, I've got the trial version of Informix, as well, which IBM now owns. I believe Informix for Linux has been totally pulled - I'm not even sure if they still sell Informix at all, given their efforts to push DB/2. I don't know if anyone high enough up in IBM reads Slashdot (they might - IBM has been Linux-friendly for a while now) but I'm sure there is still code from dead product lines of that kind that IBM could "recycle" into the Open Source community.

My guess is that IBM is reacting, to some extent, to Ingres being released as Open Source, as the prior Open Source databases (Postgres and MySQL) are not generally regarded as "enterprise ready" or usable at the "data warehouse" level, whereas Ingres was. IBM would undoubtably prefer people to be using their database (even if at zero cost) than a competitor. I'm not sure how many tools people developed for Ingres as a result of it being Open Sourced, but CA claimed a success. As IBM are also targeting developers, it seems reasonable to suppose that they are assuming that they can achieve similar results.

Re:It's a trial copy

[AKAImBatman]

My guess is that there's a bug in the timelock. I don't recall seeing the trial version available for a while, so my guess is that it got pulled, but I couldn't swear to that.

Nope, I'm talking about the Personal Development Edition. (I looked it up. ;-)) The PDE is absolutely free for development use, and contains all the major functionality. IIRC, it has similar hardware limitations, but that's not something you really notice during development.

Re:It's a trial copy

[Forbman]

IBM is still releasing updates to Informix (just worked at a now-Oracle shop that had just upgraded Informix versions).

Re:It's a trial copy

[Shade]

Actually, the recently-released IBM Informix DS v.10 now is IBM's FLAGSHIP database software for pure OLTP applications. We have been an Informix shop for nearly six years now and as many others we were pretty concerned about IBM's plans for IDS, the day after the acquisition. Fortunately, I can assure you that it is quite alive and kicking, and only gets better and better. Fine, fine system.

IBM Undercutting Itself?

[Zerbs]

The specs for the free version seem high compared to the Express versions offered by MicroSoft and Oracle. I've seen a number of companies run their production databases on simmilar or even lowwer specs than that. (2 dual-core processor, 4GB RAM) I know IBM is moving to the service model, and that they tend to charge an arm and a leg, but I wonder how long these companies will be able to give things away for free. Will there eventually be escalation on what is free from Oracle and MicroSoft too?

Re:IBM Undercutting Itself?

[Kadin2048]

My theory on why the specs are so high is because IBM can afford to do that, without cutting into much of their marketshare. At least on the hardware end -- IBM does offer some lower-end servers, but their bread and butter are the high-end ones. So where somebody like Microsoft has to limit the specs on the free version pretty severely, because otherwise they won't have a retail product to sell, IBM knows that a lot of people want to run their database on big iron. And big iron is a lot of money, not just in the hardware, but also in the service contracts and stuff that go along with it.

If your database is mostly used on commodity, low-end hardware, you can't give away a version that runs on a quad-core, 4GB machine: that's eating into your home market. IBM can, because a lot of their revenue (I'm guessing) comes from machines much further up the specification ladder than that. In fact, they would love you to run DB/2 on a high end machine, because the sooner you do, the sooner you'll make use of it, and probably the sooner you'll find its limits. (Following the general rule that software expands to fill whatever resources you allocate to it.) And when you hit the limits of the commodity/low-end hardware, IBM would be more than happy to help you migrate your DB/2 install into something a little sweeter. For a price, naturally.

Also, since they're last to the free-version game, they want to one-up everyone else. Simple competition.

Anyway, I think the "spec creep" is a good thing for consumers, both IBM's and otherwise, because it might cause a 'free version war,' that can only be a good thing in the end.

Re:IBM Undercutting Itself?

[Tekfactory]

You aren't getting the IBM strategy, its the same reason as IBM got behind Java a few years back. They support Java all the way up to the S/360 Mainframes.

Free to developers means they are more likely to use it on paying gigs.

BUT free to commercial use means.

You can play for free with DB2 on small hardware, run a small business database on it, and when you need to expand your business, you buy real DB2 no recoding required, and they will be happy to sell you Enterprise Servers, or Big Iron to run it on like an RS 6000, or IBM S/360. And Services, and Support.

Re:No, mysql does not jail you with proprietarisms

[RetroGeek]

proprietary extensions

select * from mytable limit 4,10

Amazing newgroup support

[Anonymous Coward]

I've been using DB2 for several years and I have to say that one of their strengths has to be the active and helpful user community that posts to comp.databases.ibm-db2.

Compare the quality of information provided ther to any other database group or forum and I think you'll be impressed.

Re:Amazing newgroup support

[Astroboy!]

The Express-C product has a forum already set up through developerWorks

Re:Amazing newgroup support

[alder]

Not to mention the great DB2 SQL Cookbooks.

Do not generalize

[alexborges]

I mostly agree with the parent's comment except for:


First of all, the commercial database offerings are far more feature rich than MySQL, though MySQL is getting better all the time. MySQL is fabulous for the hobbyist and small business crowd, but won't be found as a core technology of a major investment bank or backing an SAP installation at a large manufacturer anytime soon.


And i dont agree because MySQL IS found in major operations of major businesses all over the world.

Okay, so perhaps not on evey single critical task for which other (both free and otherwise) databases excel better, but for simple relational, transactional applications, its a pretty damned good database. And its free and simple and has a huge community arround it.

Yeah, so we have no 'native' XML support on it, but is that really critical for every application? NO! Normal and critical client/server apps have been working well for ages on stupidly bad db engines (fox, access), whats wrong with mysql then?

Re:Do not generalize

[RevMike]

And i dont agree because MySQL IS found in major operations of major businesses all over the world.

Okay, so perhaps not on evey single critical task for which other (both free and otherwise) databases excel better, but for simple relational, transactional applications, its a pretty damned good database. And its free and simple and has a huge community arround it.

Yeah, so we have no 'native' XML support on it, but is that really critical for every application? NO! Normal and critical client/server apps have been working well for ages on stupidly bad db engines (fox, access), whats wrong with mysql then?

I'm not talking about things like XML support. I'm talking about things like views, triggers, stored procedures, a PL/SQL or T-SQL like language (I know that these objections changed recently/are changing). These are huge issues for lots of clients. I'm not saying that you are not going to find MySQL at big enterprises, but that MySQL is often a niche player running a certain bit of infrastructure, backing a corporate website, behind DNS or LDAP, things like that. MySQL is only now getting features that Oracle/Sybase/DB2 were offering 10 or 15 years ago.

The fact that you compare MySQL to fox and access is telling. MySQL is, of course, light years ahead of those products. But the releases of DB2 and Oracle in production (which are roughly two years old now) are a similar distance ahead of MySQL. MySQL is closing the gap, however, and I expect it to be roughly on the same level as MS SqlServer in by 2008, and a potential threat to Oracle and DB2 at that time forward.

Link to IBM's product page

[GrandWaz00]

DB2 Express-C for Linux and Windows

So what?

[Qbertino]

I don't want a free database. There are more full-range free databases out there than free full-range editors.

I want a free database + free zero hassle one stop installation + free zero database driver suckage + free native object-relational OSS PL support (and I mean Python and Ruby and PHP and Perl, all at the same time) + free full range plattform independent grafical admining + SQL errors that don't say "syntax error between line 3 and 10000" + a free full-range professional level grafical ER tool with reverse engineering of any DB I have to migrate to the DB they offer + free optional zero fuss, 3 config lines maximum load balancing.

Call me when you offer that and I'll be using DB2, Oracle or whatever within an instant.

Until then I'll stick to my current MySQL InnoDB stunts and my plans to migrate to Postgres or - as the case might be - Firebird. They are truly free and they got my attention. And if MySQL Workbench will be as cool as it looks I might even just not switch at all. Despite the fact that current MySQL still has way to go before becoming a full range database. MaxDB might change that - but we'll see.

Sorry folks but getting attention requires a tad more than just giving your DB away for free these days. And it's all your fault. Hadn't you asked such bizar prices in the first place things probably would look different today.

Re:So what?

[Bozdune]

I think the parent post says it all. Tell me something I haven't heard before, don't scare me with limitations that will probably bite me in the ass later or stick me with a surprise licensing bill, make it worth my while, make it easy to convert over. Otherwise, I'll stick to what I know, whatever that happens to be, because I can be productive there.

Case in point: Thunderbird. Latest version still won't import all my Eudora mailboxes, or retain the notations associated with them (like, note that there's an associated reply). What am I supposed to do, run two mail clients, Thunderbird for "new" mail, and Eudora so I can look up "old" mail? Doesn't take a genius to figure out that continuing to use Eudora (which I despise) is still a better answer than converting to T-bird.

Re:So what?

[Per Wigren]

Search for Eudora on and you'll find a whole bunch of tools to convert your Eudora mail to standard mbox format which just about any mail program (including Thunderbird) can import...

Re:So what?

[Bozdune]

Thanks, but apparently I have to hack someone's long-abandoned Perl scripts or Python scripts (support was dropped for the two relevant projects 4+ years ago) to accomplish the conversion. Instead I think I will do something that may make me richer.

Pretty easy decision.

Re:So what?

[Anonymous Coward]

But... but... you read Slashdot! Why wouldn't you want to sit around and code your own fix for EVERY STINKIN' IT NEED?!

Re:So what?

[Uncle Kadigan]

There are more full-range free databases out there than free full-range editors.

But are there free-range full databases? And do they taste like chicken?

Linux on IBM hardware

[toby]

when will you offer either a 'naked' Thinkpad, or one that comes preinstalled with Linux (or FreeBSD) for us who want a real workstation

Well, you can order IBM's servers with either SuSE Linux Enterprise Server or RHEL. Their workstations can be ordered with RHEL.

For ThinkPad you'd need to visit Lenovo, and sadly it seems you're right: So far Microcrud only. They've even sold their soul Dell-like to the extent of "recommending" it on the home page. Send them a complaint! :)

Scalability and windows commentary

[finnif]

DB2's XML functionality looks awesome, I'm definitely going to try this out.

I am interested in all of these free editions for my web projects because it seems like it offers simpler scalability than MySQL down the road. Most high-end web developers--which I certainly am not--often end up having to find ways to get off of MySql as the site grows. Am I the only one who likes the idea of these scalable databases that I can buy later on?

On the other hand, if you're on Windows, it seems like MSSQL 2005 is your clear choice. It surprises me that any of these companies are bothering to compete against Microsoft there. It's beautifully integrated into the Visual Studio IDE, as well as offers .NET-compiled stored procedures and other Windows-centric friendliness. Oracle and IBM have to come up with great linux tools to compete against what Microsoft has created in MSSQL 2005 (Oracle's Raptor is a start)

Ooooooo....

[Anonymous Coward]

Can I get a free AS/400 to go with it???

Re:Ooooooo....

[Anonymous Coward]

Can I get a free AS/400 to go with it???

At first I thought about moderating this coward up as funny, but after two months working with DB2 on an AS/400, I can tell you there's nothing funny about it at all!

Open source it...

[m50d]

and tell Scott McNealy what to do with his assets. 'cos we could really do with java 1.5 being free.

Why the hell not?

[LWATCDR]

If you are a GPL or death person then there is not reason. I would rather us Postgres than MySQL but that is just me.
If MySQL works for you and you don't need anything else and don't want to learn anything keep MySQL.
Now for the reason you would.
Want to learn how to use DB2?
Want a really heavy duty ACID compliant Database server?
The limits on DB2 are... TINY I mean even in a good sized company a 4 gig ram limit and two core limit is unlikely to cause you issues.

This is to target MS-SQL not MySQL. If you develop for MS-SQL you are not only locked into MS-SQL but also Windows. DB2 Runs on everything under the sun including all the big iron IBM sells.
Good move for IBM. I may even install it to take a look. I will probably stick with Postgres since IBM could always make it not so free in the future. But if I ever did need to scale it one of my projects to Enterprise size, DB2 is a good option.

Re:Why the hell not?

[LWATCDR]

Man I can not freaking type today at all.
use not us, no not not.

Wow. No user limits or db size limits?

[WoTG]

You can serve quite a few users on a 4 core server! To me, this looks like a direct attack on MS SQL Server. A lot of software for small and medium sized businesses run on SQL Server. I doubt that IBM has much sales volume at the low end anyway, so what have they got to lose?

Ummm ok so where is the Free Beer?

[Dream1979]

Oh man you mean its BYOB?!?!?!?

Don't forget U2

[Larry Lightbulb]

There's been free personal editions of UniData and UniVerse available for years -

"...As of UniVerse 10 and UniData 6, the Personal Edition is available as a free, downloadable, 2-user license with some functional limitations. Personal Edition is available for 32-bit Intel x86 servers with one to two processors running Microsoft Windows or RedHat Linux. There are no maintenance offerings for the Personal Edition. This edition is designed for single-user personal development and training as well as sales demos. It is not for remarketing or use in production environments..."

Much easier than any other database out there.

I don't pay for software until I know it well.

[pkulak]

For example: I'd heard good things about ASP.NET 2.0. I've been a Java guy for a couple years, but decided to try it out on a new project. I got the VWD package for free, and got going. After a couple weeks I realized that this was the direction I wanted to take with future development and got my boss to order up some Visual Studio Pro licenses for myself and coworkers.

I started in Java because it was free, and would still be there had it not been for VWD. I think this is an excellent strategy by IBM.

Re:I don't pay for software until I know it well.

[plague3106]

As an aside, there are free .Net IDE's out there (well, one at least). Google for SharpDevelop.. if fact the UI looks very similar to Studio 2005.

The site that has SharpDevelop also has some other useful .Net components, like SharpZipLib, which allows you to work with Zip files.

thankfully, not

[penguin-collective]

Native XML capabilities don't belong in a relational database. If you want an XML database, use one.

Re:thankfully, not

[TheRaven64]

I have seen a lot of things recently that try to shoe-horn hierarchical data (usually XML) into SQL databases. Perhaps you can suggest a more optimal solution? I'm sure one must exist, but I've never been motivated enough to look.

Re:thankfully, not

[Anonymous Coward]

As I was saying: don't try to shoe-horn, just use the right database for the kind of data you have.

However, I have to say that generally I find people overuse hierarchical structures; they are usually chosen more for convenience of the implementor than actually being a good data model.

MySql Is not free (beer)

[aled]

Alright, I run a MYSQL database at home for my little crappy web server which no one visits because pictures of me visiting Mexico isn't exactly worth typing a number into your browser. All those details aside, why should I switch to either Microsoft's SQL Server 2005, Oracle 10g or DB2?

No reason I guess, unless your site isn't built on open source software, on which case you should be paying your commercial MySql licence.

From http://www.mysql.com/company/legal/licensing/:

In their simplest form, the following are general licensing guidelines:

• 
  
• If your software is licensed under either the GPL-compatible Free Software License as defined by the Free Software Foundation or approved by OSI, then use our GPL licensed version.
• If you distribute a proprietary application in any way, and you are not licensing and distributing your source code under GPL, you need to purchase a commercial license of MySQL
• If you are unsure, we recommend that you buy our cost effective commercial licenses. That is the safest solution. Licensing questions can submitted online for our advice, and we encourage you to refer to the Free Software Foundation or a lawyer as appropriate.

Re:MySql Is not free (beer)

[3770]

I believe the GPL only applies to software that you are distributing right?

The rule is, if you distribute the binaries, you must also distribute the source code. But if you only run your code on your web server you don't have to do that.

Have I misunderstood something?

Re:MySql Is not free (beer)

[Software]

>No reason I guess, unless your site isn't built on open source software, on which case you should be paying your commercial MySql licence.

I think you're misinterpreting the MySQL license, which is the GPL, as interpreted by MySQL AB. Whether or not eldavojohn's site is built on open-source software is irrelevant because he's not distributing the software. He's only distributing HTML files, so he's under no obligation to purchase a license for MySQL. In fact, the link above specifically exempts eldavojohn from having to release any code:

Free use for those who never copy, modify or distribute. As long as you never distribute the MySQL Software in any way, you are free to use it for powering your application, irrespective of whether your application is under GPL license or not.

This is sometimes called the the "Application Service Provider" loophole in the GPL (since MySQL specifically mentions that the GPL doesn't apply in this case, it's not a loophole, but some people call it one anyway). ASPs don't have to publish their modifications to GPLed code, because they don't distribute the code. They only distribute the "output" of the code, which is not protected by the GPL.

Re:MySql Is not free (beer)

[Anonymous Coward]

This is sometimes called the the "Application Service Provider" loophole in the GPL (since MySQL specifically mentions that the GPL doesn't apply in this case, it's not a loophole, but some people call it one anyway)

Then explain why MySql released the FLOSS exception list?

All MySql Client libraries for C, Python, PHP are licensed under the GPL, making any closed-source website application using these libraries unpossible, right?

I've been trying to research this annoying license issues for a python-based website

Transactions and Commit functions

[Kadin2048]

I keep hearing rumors/statements that MySQL's support for transactions and commits is lacking or "is a hack." It's not really my area, and I've played around with it only on a very superficial level on a home server, but in its default configuration it did not seem to do commits, but rather changed elements of the database in real time as it was being given the data, with the possibility of a half-created record being left in the system if (say) the client got disconnected halfway through. This would make it unsuitable for any kind of transaction-based processing, where the integrity of what was in the database was paramount.

Can anyone who's used MySQL respond to this? I'm sure there has to be a way to do it, but it gets spread around quite a bit by people who want to write MySQL off as a 'hobbyist' platform.

Re:Transactions and Commit functions

[alexborges]

MySQL has a fully transactional engine that does the log thing, the rollback thing and the rollforward thing. This engine is ACID compliant from any way you wanna look at it.

The trick is the table format you choose. Refer to the mysql docs for that.

Re:Transactions and Commit functions

[rodent]

Full ACID compliance can be found when using the InnoDB table type. Just as with any DB, you will lose speed compared to MyISAM but you can do rollbacks as with any other real DB.

That being said, if you're going to lose speed, try Postgres. It's superior to MySQL in many ways and makes a much better trasitional path to Oracle/DB2/SQL Server.

Re:Transactions and Commit functions

[Jamesday]

The *nix default database storage engine is MyISAM. If you're using Windows and tell the installer you want transactions the default is set to InnoDB. The transactional storage engines include InnoDB, Cluster and BDB. To use one of those either change the default database engine or use engine=InnoDB (or whatever) in the CREATE TABLE statement. You can convert from one to the other using ALTER TABLE ENGINE=InnoDB or whatever you want.

Also, the default transaction mode is autocommit, automatic commiting after each statement. So if you don't either start a transaction or set autocommit=0 you're going to see automatic per-statement commits.

See Transactions in MySQL and transaction isolation levels for more.

It's been this way since before MySQL 4.0, though using "type" instead of "engine" in versions 4.0 and below.

Ah!

[jd]

Ok, I had the trial Universal edition. If someone patched KDevelop to support DB/2, would that make it an IDE for KDE for PDE?

Re:Ah!

[AKAImBatman]

If someone patched KDevelop to support DB/2, would that make it an IDE for KDE for PDE?

Ouch. That makes my head hurt.

This move by IBM was predicted

[Anonymous Coward]

Mark Whitehorn , writing in "The Register" producted a move like this back on 18 Jan 2006. He seems to be a real fan of DB/2 -- see "DB2 - the secret database" at http://www.channelregister.co.uk/2006/01/18/db2_ne glected/ My feeling is that IBM is finally getting it that developers like to use what they are familiar with, and what they become familiar with is what they can use on their own for small and/or home-brew projects. (Free beer anyone?)

WTF...

[Tony Hoyle]

Oh my god.. you killed IBM!!

The download just aint happening right now.

This relase good news.. with any luck I can write a db2 plugin for the next release of my software (if only SQL was a real standard and you could support a database with just ODBC... sigh...)

Re:Wow. No user limits or db size limits?

[Zemplar]

A 4-core server? Are you certain that the "2 processor" limit excludes multi-core CPUs? I wouldn't think so.

Why not

[AgentCharlieBrown]

The question is not why but rather why not? :O)

Where's the OS X version?

[jocknerd]

I see the Linux PowerPC 64 bit version, but how come there's no OS X version? Come on IBM, get with the times!

Re:Where's the OS X version?

[C_Kode]

I doubt that will happen for a while. At least not until they fix all the performance issues with the Mach kernel. I just wouldn't use OS X for any type of heavy DB (of any app other of the same nature) use until then.

A cheap file server with X-Serve. Sure!

Re:Where's the OS X version?

[egroeggnik]

I see the Linux PowerPC 64 bit version, but how come there's no OS X version? Come on IBM, get with the times! Umm. Maybe they ARE with the times. I haven't seen alot of Mac servers in production. I know they make them, but I have yet to actually see one in use.

Re:Where's the OS X version?

[TheRaven64]

OS X 10.4 makes a nice development platform. They have done the 'make it work' part quite well - all of the bits of POSIX I've tried to play with (including aio and pthreads) work nicely. What they haven't done yet is the 'make it fast' bit. Thread locking operations, for example, completely kill performance. All system calls cost about ten times as much as they do on other platforms, so anything that is I/O intensive is going to cause problems.

Oh, and the VM subsystem is horribly, horribly, broken. I have no idea what's wrong with it, but it spends far, far too much time thrashing. I have never seen an OS quite this bad (well, maybe Windows, but certainly nothing UNIX-like. FreeBSD, which uses a Mach-like VM system similar to the one OS X inherited is orders of magnitude better).

It's for consultants

[shummer_mc]

In this world of outsourcing, companies want to buy "enterprise" software packages, experts, and "solutions" (whatever the hell that means) for a cheap price from a local company (IMO- mainly to avoid responsibility for potential failure). If these tools are not free, then there are no experts/packages/solutions avaialable because these smaller companies cannot afford to buy "enterprise" databases... enter: free licenses sized for consultant shops.

As for the post asking why these solutions are preferred to the OSS varieties, it's really pretty simple: Large companies won't spend money on non-branded enterprise software. I used to work on an app written in Sybase's ASA (non-enterprise). Well, it worked really well. For the amount of data and the amount of processing required for the app, it was sweet! No DBA required. Anyhow, we went to sell it to other customers (large gov't contractors) and the first Q I got was... Does it work with Oracle? We ended up converting it to work with SS2000 (this was for a large company in Canada, actually) because they wouldn't allow ASA on their servers. It's dumb, IMO.

Most OLTP databases would be much more efficiently done in 'lighter' databases. OLAP... sure, do it in one of the big databases.

Of course, I'm not a DBA...

Re:It's for consultants

[Forbman]

OLAP would be better served by "lighter" databases, not OLTP.

Re:It's for consultants

[shummer_mc]

Really? I'm curious as to why you think so. My reasoning was as follows:

OLTP- online transaction processing... (light datasets, mostly inserts/updates, a bit of processing)
OLAP- online analytical processing... (cubes, stars, and other random geometrical shapes;) HUGE datasets)...

I'll give you that I've not worked on a truly OLAP system, but I get the impression that the processing end is fairly heavy, given that (some) analytical data needs to be derived (not to mention that some data-transformation usually needs to occur to move OLTP type stuff into the OLAP datastructure). Also, given the rather large datasets (usually historic data from an OLTP system) these are prime targets for technology like partitioning, etc.

I'm not a DBA, like I said, but why did you think the opposite?

Re:It's for consultants

[kpharmer]

> OLAP would be better served by "lighter" databases, not OLTP.

What gives you that impression?

Consider that (R)OLAP doesn't benefit much from btree indexes - which are ineffectual for selectivity of greater than 2% of your data. That is, if you want to analyze 100% of the data for a single month in a database with 12 months of data @ 1,000,000 rows per month - a btree index won't be useful in just selecting one month of data. Partitioning will be useful - but you're only going to get that from db2 or oracle (or on sql server's immature implementation at ~$30k/cpu now). Light databases won't support it. Right off the bat you just took a 12x performance penalty by going with the lightweight product.

And when postgresql/mysql scan that 12,000,000 rows - they're doing it single-threaded. Even if you're on a four-way server. Whereas db2 or oracle can parallelize that query - giving almost a linear performance improvement. That's another 4x performance penalty you took by going with a lightweight product.

In other words, db2 or oracle would have been about 48x the speed of postgresql or oracle in this common olap scenario. And that doesn't even get to optimizer maturity (will bury mysql), memory tuning, OLAP functionality, materialized views, etc, etc, etc.

Some day the lightweight products will get there. Of course they won't be so lightweight when they do ;-).

Oracle on PC hardware, Windows vs Linux

[KenSeymour]

One thing I wish I understood more is why the same desktop machine can run Oracle nicely under Windows, but you can't even install it under Linux.

The machine I have in mind has 512MB of memory, some Pentium 4 processor and the usual 40-80 Gb of disk.

At work, we use a plain old Windows desktop machine for the development Oracle server and don't have any problems.
I tried to setup Oracle under Linux and it would require more memory and lots of unusual System V shared memory conifuration and Kernel settings etc.
The Oracle install under Windows runs a GUI based installation. You click through
a bunch of stuff and boom, you have your instance up and running.

Our production environment is a pair of Unix servers using Oracle multi-master replication.
When we have to bring both servers down and run Oracle, JBoss, and our application server
all on a windows Laptop.
Does Unix/Linux make it harder than it needs to be?

Re:Oracle on PC hardware, Windows vs Linux

[jedidiah]

No.

If you can't manage setting shmmax on an Oracle database, you really should running something a little more shiny and happy. If you are complaining about shared memory parameters (which are all VERY well documented by the included documentation) then you really shouldn't be running Oracle since it's own parameters make the stuff in sysctl.conf look like a cakewalk.

With few exceptions, Oracle exposes THE EXACT SAME interfaces to all operating systems including Windows. The same crappy java or j2ee tools are going to be present under any OS. Ditto the commandline tools.

The same hidden and non-hidden init parameters too.

Oracle is by no means a self-operating product.

Re:Oracle on PC hardware, Windows vs Linux

[KenSeymour]

My Oracle installation experiences have included:

1) Compaq Tru64 Unix
2) Windows
3) Linux

Install 1) was the worst. There were so many patches and bug fixes for it, it took several days.
It was so difficult, we used rcp to copy the disk image onto later servers to install it.
Then we run the instance creation scripts.

For some strange reason, install 2) is incredibly painless.
I know from experience that Windows IPC calls are fairly different from Unix (System V) IPC calls.
Since Oracle on windows does not require the cygwin libraries, I will assume Oracle has a lot of stuff like:

#ifdef __WINDOWS__ /* do it one way */

#endif

#ifdef __UNIX_flavor_1 /* do it another way */

#endif

Install 3) was at home, so I did not have several full days to devote to it. IIRC, the required shared memory area size was bigger than my total system memory. So shortly after
that, I gave up.

We used a couple of DBA consultant to help us with the Tru64 Unix install to help us with the Oracle "hidden parameters."
But my point is, if you run Oracle under Windows, you don't have to fool with any of that stuff.

The lesson I walk away from that project is this:

If you need to run Oracle on Unix or Linux, make sure you have a good Oracle DBA, preferably full time.
If you can run Oracle on Windows, you don't need the DBA.

What I haven't tried yet is to run Multi-master replication on a pair of Windows servers.

I subsequent project I am working on uses PostgreSQL and it seems pretty painless to use.
I have used that both on Windows and Linux and have not had any difficulties.
The client for that job has a mixed Windows and Linux environment.
I am not aware of them having a PostgreSQL DBA, either full time or part time.

They Should Generate Big $$ For Services

[Black-Man]

Since UDB is a piece of crap! This is a ruse, big time. And no... I don't work for Oracle - I just use both.

WT..:.?

[Anonymous Coward]

What is so enterpise about this crap?

and wtf is $this->button1 = new klorofil_gambart_component_Button($this); ?
is it really nececarry to type klorofil-crpa-more-crap-andmore.button?"?!

maybe for fun, probably not for work

[protocoldroid]

I once worked for a company, we were using DB2 EEE. We were developing an application for proof of concept, and... Installing DB2 (across multiple machines for paralell processing) was a little bit like having my teeth pulled (i'd say "compared to MySQL" but, i've never used mysql in paralell -- if you can [i'm not sure]). But when that was done, I was quite impressed with DB2. We were though, working with very large databases. I really became quite a fan of it, and my eyes opened wide when i saw this topic headline.

Since then I've stopped "working for the man" and am a self employed developer. What's my database of choice? MySQL. I wasn't involved in paying for DB2 EEE, but, I did hear a rumor around the water cooler that it costed $20k / Year / Processor. Which, as a small business person is quite a bit out of reach. After my good experiences with DB2, I would consider giving this a whirl. In fact when i first saw this topic I thought "Yes! I get DB2 for free!" then I started thinking "Holy guacamole... reworking code for days/weeks on end to interface with a different DB system". Plus... I don't work with really huge databases anymore.

I once heard an urban legend that MySQL is faster than DB2 with queries on tables that are less than 250k rows. If that's the case... It's really a rare table for me these days that has 250k+ rows (and when I do have that many records in a table, it's usually not data that i need access to quickly, but for historical puposes).

I will however, give it a try with some small (more-or-less-for-fun) projects. Being self employed, I'm not so worried about having the specific "resume points" from using it, but... More or less curious to give it another try.

One of the most impressive additions to DB2 that I know of is the DB2 Spatial Extender, which "[adds to DB2] a set of advanced spatial data types that represent geometries such as points, lines, and polygons and many functions and features that interoperate with those new data types." I believe it's marketed towards and tailored to GIS people, but I had fun experimenting with it while developing a project that needed to store 3d cartesian coordinates.

Re:maybe for fun, probably not for work

[kpharmer]

> I did hear a rumor around the water cooler that it costed $20k / Year / Processor.

No, it could cost $32k/cpu list plus something like 15%/year for maintenance. Of course, that's if you want a beowulf-cluster like solution.

You could just run the workstation edition on a single 4-way and still support vast amounts of data. At between $1500 & $7500 / CPU list that can easily handle a data warehouse with billions of rows of data - using MDC, summary tables, etc, etc. Note that db2 (and Oracle) can be often be cheaper than MySQL on smaller systems.

> I once heard an urban legend that MySQL is faster than DB2 with queries on tables that are less than 250k rows.

Could be - mysql could be faster than db2 on extremely small hardware. Then again, db2 has better memory tuning - so you could probably tune db2 smaller. Not positive of that. And mysql won't be faster on writes and certainly not on large selects on a SMP.

> Installing DB2 (across multiple machines for paralell processing) was a little bit like having my teeth pulled

Yep, until a few years ago all the databases were pretty nasty to install. SQL Server was the first one that was really easy - around 2000. It took Oracle and IBM years to clean up their installs. They're pretty easy now though, wizard-driven if you want it, other wise just a few simple commands. EEE is still more than that - since you're installing it on sometimes dozens of separate machines, but it's also easier.

DB2 is SOLID

[wardk]

May not be as light as MySQL, nor the darling of the open source community.

But don't kid yourself into thinking DB2 is nothing less that what it is. a ROCK SOLID, HIGHLY SCALABLE, MATURE database.

how useful all the additional features might be to the average PHP developer is certainly questionalble, but for the serious developer looking to do serious work, this is an excellent addition to the available toolkit.

  this is good news indeed, thank you IBM.

(OT) Re:So what?

[Anonymous Coward]

[My english is better than most other people's german, so please point out mistakes politely. Thank you.]

grafical => graphical

Most people don't need this.

[ChrisA90278]

Most people who just need to get to work in the morning don't need to buy an 18 wheel semitruck and trailer. Heck, most don't even need an SUV. So if freightliner (who makes the big trucks) started giving them away most people wouldn't even want one Same here. DB2 is not something your avaerage home computer user would want. It is something of an 18-wheeler of a DBMS. Something only someone with a really big job would need. It's not surprizing that for most uses mysql works. There are far more users with small and simple neds then large 1000+ emplyee enterprizes. What you gain by using the likes of DB2, Oracle and maybe even POstgresql is "scalablility". The ability to handle larger demands by adding hardware. Notice that IBM be limiting use of the free version to machines with "only four CPU cores and 4GB RAM figures that the free version wil not compete witht ehothr version. So the "real users" of DB2 are using much more powerful servrers.

One word...OS/2

[dtjohnson]

IBM stopped selling their OS/2 operating system last year but has refused to release it as open source or even to allow 'free' licenses for additional users. If IBM cannot find the generosity to release a 'free' version of a defunct operating system which there are cash-paying former customers who could actually benefit, there's reason to be suspicious of a 'free' version of DB2. The DB2 free version is probably some sort of scheme for getting lock-in to the platform and is not just an example of generosity by IBM.

Re:One word...OS/2

[kpharmer]

> If IBM cannot find the generosity to release a 'free' version of a defunct operating system

OS/2 is rumoured to use third-party software licensed by IBM from Microsoft - that they can't open source. How does this make them "not generous"?

> The DB2 free version is probably some sort of scheme for getting lock-in to the platform and is not just an example of generosity by IBM.

Lock-in? Well, getting mindshare at least. Which puts them in the same company as:
      - sql server (free extremely limited database version)
      - oracle (free extremely limited database version)
      - mysql (free vaguely-defined database version)

The only database products that I'm aware of that don't do this are Postgresql and Firebird. Neither of which match db2 for performance. And nor did this free version of db2 just suddenly appear - they had a very low-cost version for a few years.

Re:One word...OS/2

[Anonymous Coward]

I would never again use IBM software for development of mission critical systems. It's not about the quality, that was allways great even when setup and configuration of IBM software is sometimes combersome. It's about long term reliability. IBM's core business has allways been hardware (and now also services), software is just a appendix. At any day another new manager may walk into the chief office and drop DB2 snipping with two fingers, just the same way it happened with OS/2 and also VisualAge C++ to name two IBM products we had built software for in the past and found ourselve at a dead end over night.

For mission critical systems which need to run and be maintained many years, open source products or companies like oracle or trolltech who's core business is the product you are using are much more reliable. They wont drop thos products just because of a managers whim.

Re:One word...OS/2

[wilec]

Hi D, I have to call you the issue of IBM releasing the code as I am sure you do not know the legal restrictions they probably face. Also I somewhat understand the reluctance of IBM exec's to push OS/2 too hard considering the legal trouncing they had just taken over monopolistic practices. But then they got less than they deserved there as well. The legal trickery and patent manipulations employed by IBM management, and complient judgicial interests, out of greed and fear of competition have greatly restricted development and set some horrible precedents.

However I do agree that generally IBM is not to be trusted. The way they misled both OS/2 users and developers was as bad as anything Microsoft has been charged with. They should be required to reinburse many customers and developers with the settlement they got from Microsoft over the OS/2 debacle.

Heck at this point I would get plum silly over access to USB 2.0 drivers for the cost of Microsofts latest OS "upgrade". I hate it but I do believe IBM has finally delivered a fatal blow to OS/2. Heck these days even the legendary OS/2 USENET groups are mostly flame wars over eCS. That is why this is being posted from a Fedora Core 4 install. I just hope they don't find a way to muck up the Linux world too.

Please note this is not intended to disparage the gifted folks who developed OS/2 and other IBM technologys or many wonderful IBM managers and staffers that tried to do the right and what I think would have been the smart thing. However overall I for one will never put any serious percentage of my eggs in one of IBM's baskets again. This is a shame because they do have, or in some cases had, some of they best developers and engineers in the world. Piss poor top level management, mostly clueless marketing execs, an ethically challanged legal staff but great engineering and science, shame, shame, shame.

Matthew

Sounds a lot like SQL Express to me...

[Anonymous Coward]

http://www.microsoft.com/sql/editions/express/defa ult.mspx

Different Needs

[raftpeople]

"I'd never want the FULL version of these databases, even if those, too, were free"

Well, in an enterprise application environment you would want them. In an enterprise application environment, footprint is near the bottom of the list of priorities. The top of the list include:
Reliability Scalability Performance

MYSQL Optimization

[Dareth]

This is SO ironic... as I have had a MYSQL query "loading" for the last 15-20 minutes in my other tab. I am getting the feeling it is not going to finish anytime soon.

Re:MYSQL Optimization

[zardo]

Yeah I experienced that too, thats because of the lousy query optimizer. You can even manually specify indexes and it will still not use them when you want it to. I don't know the explanation for that, but "lousy query optimizer" pretty much encompasses it doesn't it?

Of course it is. Do the logic

[Flying pig]

IBM does not do much with SMEs. But some SMEs will become large enterprises, and when they do they will want big tin and support and they will want to keep on using what they've got.

The logic is exemplary. Microsoft has a lot of SME installations of MSSQL, and they hope that some will become big installations and by then they will have figured out how to do big iron. So IBM says to the SME "come to us instead and no matter how fast you grow your investment in software will never be obsolete. Hey, worried about viruses and licensing and stuff? Want to start off from day one with Linux on the server and the desktop? We can do that for you."

The reasoning is sufficient to have caused me to download DB2 for Linux to install on Ubuntu tomorrow morning, if only to evaluate just how difficult a port will be for us.

Re:Of course it is. Do the logic

[swordfishBob]

.. and have you noticed how easy it is to legally have and use MSDE, which is SQL Server with a size limit and without the GUI? That has to be a significant funnel of customers for MS, and a real nuisance for their commercial competitors.

I agree completely -- wish I had mod points now...

[CFD339]

IBM is effectively saying "If your product is targetted at x86 servers -- even multiprocessor/multi-core -- you don't really have a product."

Microsoft's market is exclusively (or nearly so) now forced to compete against a product that most of the industry considers vastly superior, and yet is free to anyone using it on a scale that Microsoft typically sells to.

IBM Isn't, and has never been effective at selling to SOHO. By giving away this market, they have the opportunity to capture some of it, poison the water supply of their competitor in Microsoft SQL, and for those successful companies that grow from PC based servers to serious hardware they can easily migrate them to iSeries or zSeries big boxes with lots of zeros on the price tag.

Its a solid strategy and a good product.

I remember doing a side by side comparison of Oracle to DB2 for a service company planning to scale up around 1999. I fully expected to pick Oracle -- but they were so difficult to deal with, that with no real advantage you'd be stupid to go with them. At the same time, no matter what oddball "what-if" question we came up with for a future direction IBM had a product integration path to handle the need.

In short, DB2 was a more mature, more powerful, and more extensible tool. At the time, MS Sql wasn't even in the running. Today's MSSQL is much stronger, but still not in the league of DB2 and Oracle.

Re:features--I'm of a different audience...WYSIWYG

[davidsyes]

I read that and got a headache just LOOKING at all that syntax. I'm a simpleton, relatively speaking. I have used ms access years ago, mostly in the query mode, and after having use Lotus Approach, it was was, to be charitable, quite a damn maddening experience to have even clicked around in access. I was having abscesses on my brain. I cannot even bear to touch ms access, and when I try, I get anxiety attacks. (Yeh, I know there is version 12 or so coming soon, but I am not a programmer; I am a relative simpleton.)

I USED to try to use Borland Paradox v4 & v5 years ago, but that went away when my Paradox-using Beta Programs manager introduced me to Lotus Approach. Approach is, for many, an underwhelming, not-too-powerful (and in my heart and mind, a woefully under-used) "relational" database application.

I am building a screenplay/dialog tracking system in it. I built an assets/inventory app when I was trying to start an Internet Cafe (In 2001/2002, I wrote the plan, filed papers, but never got space or loans, so it was a good exercise in 'rolling my own' database application...), and I wanted every last detail in customized layout. I had many dozends of fields for capturing data.

One thing I cannot stand in Approach is the cobbled-on "Net-It-Now" thingy which is an attempt to make HTML pages. At least that's what I take from it.

I wish IBM would open-source the code they OWN within Approach and let the community use it as a front end to databases. I wish they'd let us make or extend our Approach-based interfaces and expand them. Approach probably doesn't need XML and other stuff, but being able to roll back transactions without some of the convoluted steps necessary would be nice. The Open Source community (with user-experience from myself and other Approach users) could enhance:

-- the Create Joins dialog (make it flexible/collapsible/editable)
-- the Import Records so that the destination fields' names could be edited or added to leave the existing fields untouched
-- the repeating panels by adding a horizontal slider to the panel
-- enable conditional text formatting/coloring
-- enable the form letters to have italicized text within the form box
-- the user security
-- the user experience by allowing a stand-alone executable
-- the query view by recording and exporting and saving on a form or a hidden/callable form the queries the user created (not just record them in the power-click/title bar area
-- charting tool to make it possible to edit more aspects of the charts, like being able to specifically enlarge one pie chart in a series, but manually shrink an adjacent chart
-- charting tool by making it possible to move the charts themselves, not just radio-reposition the ancillary text
-- the form view by allowing chart components to be placed on the chart (IMAGINE the query forms showing charting information; yes, I know the reports can be turned into repeating panel reports and that a chart can be embedded on the panel, but it's quite limited...)

and more.

It has been a long time since Approach has seen any awards. It's nice to know it's an "Award-Winning Relational Database Application for Professionals....", but its tragic and almost a travesty to see that in a Lotus Notes environment (where I contract) that Lotus SmartSuite cannot even be introduced to and gain traction here. Lotus Approach on top of Notes, and Lotus Approach and Word Pro would KILL the dizzying xls-pervasion. It's disheartening to see so many spreadsheets being used with all the little cute coloring when the information should be persistent (kept) and used in DATABASES that the end-users can conquer, rather than in damned excel sheets that keep getting butchered and hammered into least-capable-user-understanding. Having excel sheets flurry back and forth is annoying when transactions could be done in the database.

I think I'm going to create myself a book on the Approach-based projects I've done for myself.

-- Screenplay
-- Trouble-ticket
--

Re:features - Postgresql replication

[rubycodez]

There is replication (master to multiple slave) for POstgresql, see slony1 , or google for commercial replication and clustering solutions for postgresql

Whoa!

[Anonymous Coward]

I'm certainly not complaining after hearing this bit of news. I've used to working with MySQL and PostgreSQL and have done a few things with Oracle too over the few years I've been developing apps. I can't say I have a preference though. I used pick the DBMS depending on the requirements of the app, rather than by my preference. My current employer, however, uses DB2 for all projects, and so far I've been quite taken with the scalability. Stress testing now makes me smile when it used to make me want to hide somewhere very, very dark and far away (That is vs. MySQL and Postgre).

The thing that has bugged me however, is that the licences are so expensive. This now means though, that I can run my own instances of the db for development and early release stuff and don't have to be bugging the DBAs for every minor detail, only when I need their expertise. Nice. Boosts my productivity.

And I do have a few personal projects going that I hope will someday be needing the scale I can get out of DB2. Now I can develop them and run them on DB2 from the start.

I know this sounds like sales pitch. Fire at will. Me happy!

MS SQL???

[justsomebody]

huh, you forgot to say dog slow. Just last year I installed MSSQL on my Linux machine and I'm still waiting for the first record in query result.

p.s. Off course real translation of my sarcasm should say, MSSQL is the only non-portable solution from named databases, which is one major thing you forgot

Re:Wow. No user limits or db size limits?

[janic]

I can vouch fo it.

Both IBM and MS consider 1 multi core CPU as a single CPU.

Thus, the Dual chip dual core machines we have are licenced for DB2 or MSSQL as dual CPU machines.

Wheeee!

Re:features--I'm of a different audience...WYSIWYG

[ckaminski]

For what it's worth, someone like yourself might be interested in Progress. Simple query language, not SQL although there is a SQL adapter available. I played with it during my short stint at Progress and I was impressed. Not bad for an embedded database, though the toolchain available for it is in some respects quite lacking.

<quote>
It's disheartening to see so many spreadsheets being used with all the little cute coloring when the information should be persistent (kept) and used in DATABASES that the end-users can conquer, rather than in damned excel sheets that keep getting butchered and hammered into least-capable-user-understanding. Having excel sheets flurry back and forth is annoying when transactions could be done in the database.
</quote>

I'd kill to be able to eliminate this phenomenon at my clients.

Re:Wow. No user limits or db size limits?

[Zemplar]

Excellent! I looked in vain for clarification on IBM's site. Thanks for the clarification.

performance and scale are related

[Stu Charlton]

But it's typically a complicated relationship. Something that performs very well with one user could very well be scalable; the issue tends to be the amount of indirection one goes through to ensure scalability.

Put another way, performance can be had by maximizing the usage of available resources. Scalability under concurrent load is achieved by minimizing resource usage, especially resources that can only be used by one user at a time.

Good software that exhibits "scalable performance" is about knowing what approach to pick, and when.

The best reason to get it (I already did)

[Mycroft_514]

I am a DBA in a DB2 shop. We are still running 7.x and they are giving away 8.2 This will give me several months to try out new features of 8.2 before the upgrade to our mainframe.

I can throw this on my laptop, and try out some ODBC stuff as well which is going to come down the pike from the development side of the house.

Free VLDB = Pay VSDB

[mbowen]

Here's an honest but perhaps offensive question. I work with financial databases and enterprise software going back a dozen years or so. In the F100 companies I work with, I have never seen any MySQL or PostgreSQL or any free databases ever used for financial transactions or data warehousing. As far as I and my professional colleagues are concerned, MySQL is a toy. Which is to say that when it comes to crunching numbers on serious iron, it's not even considered.

I admit my bias which says that anyone who programs for backends of websites is not likely to build anything that challenges a real database (DB2, Oracle, Sybase, Teradata) very much in the same way that nobody who writes for Windows is taken seriously by people who write for Unix (Solaris, AIX, HP/UX).

So I ask the following question for somebody who might be familiar with a heavy duty commerce site. Where do you make the break? What size database?

I work primarily with multidimensional databases which are even more sophisticated and complex than RDBMS, so I even tend to scoff at DB2 and stuff like Microstrategy. My basic criteria is that if it can't do the equivalent of a 10 way join on an unbalanced snowflake with subsecond response time, it's not worth my time.

Anything with less than 50 million source level records for real-time querying I consider a small database. What's large for a MySQL guy?

Re:Free VLDB = Pay VSDB

[Anonymous Coward]

Anything with less than 50 million source level records for real-time querying I consider a small database. What's large for a MySQL guy?

There was a MySQL article discussed here that mentioned a 20-billion row data warehouse. It spoke also of the shopping application for the SABRE airline reservations system that uses a MySQL server farm.
 
Neither sound like a toy.

A warning for those wanting to install it

[LordP]

You need GLIBC 2.2 to use it. If you are using GLIBC 2.3 (as I am), then you're out of luck.

DB2.EXP
      IBM_db2cliv81...
error: Failed dependencies:
                libc.so.6(GLIBC_2.2) is needed by IBM_db2cliv81-8.1.2-97 /bin/sh is needed by IBM_db2cliv81-8.1.2-97
The installation logfile can be found in /tmp/db2_install_log.15124.

db2_install program terminated prematurely.

Re:A warning for those wanting to install it

[DenialS]

True, but most Linux distributions give you the option of installing compatibility libraries, which would solve your problem nicely.

Sounds good, something to start with

[Anonymous Coward]

It is ironic that one of the very first posts mentioned a personal picture database. I have MySQL installed as part of a MythTV installation and have otherwise never worked DBs before. I was thinking of using MySQL to build a personal picture / movie database, for my personal site as well as for practice for doing a possible business app down the road. But, my business app would require scalability and *poof* this article appears. Sounds like exactly the kind of thing I need to get started with. Thanks very much IBM.

Now..what about that glibc 2.2 post? Wow...too bad.

Difference in administration tools?

[5n3ak3rp1mp]

I used and administered SQL Server for years, until Microsoft started (finally) leaving a pretty bad taste in my mouth. Unfortunately, I got used to the admin tools, which are (for the most part) nice.

I'm familiar with things like phpmyadmin and other web frontends to databases (a lesser solution, I feel). I noticed that the new MySQL 5 client tools for OS X that you get with the binary download are actually pretty awesome.

Meanwhile I don't know what the popular interfaces to Postgres or DB2 look like, and from what brief glimpse I got of Oracle, it has a whole mess of tools, emphasis on "mess".

So although I bet a lot of you DBA's are perfectly fine being terminal jockeys, what's it like actually administering these things via some sort of GUI? Etc.?

Replication support

[TheBracket]

Does anyone know if the free edition includes good replication support? I have a project that relies on asynchronous multi-master replication (Merge replication in SQL Server), and they have been looking for a 'real' database platform. They are a non-profit, and full editions are expensive even with non-profit discounts. I can't seem to find this in IBM's copious documentation on their website!

Re:Replication support

[Jamesday]

Can't tell you for DB2 but for MySQL, Wikipedia's main set of five database servers is using replication included under the free GPL license.

Re:FLOSS exemption

[Jamesday]

If you aren't redistributing the python code, you don't need a license other than GPL for MySQL. The FLOSS exemption is for those who are distributing their applications with MySQL, not for web sites. For an example reply from MySQL saying so, see Licence for MySQL use at website?.

Oracle MVCC versus PGSQL MVCC

[Anonymous Coward]

Oracle uses optimistic MVCC -- deprecated rows are overwritten sooner but the redo logs get huge with rollback transactions. PGSQL uses pesismistic MVCC -- tables get more deprecated rows and require periodic maintenance (VACUUM) but the redo logs (WAL) won't go nuts.

Oracle's method requires less maintenance to keep performance up but can suck when cancelling big update jobs. If a big update job is running, don't ever think "god I'd better cancel this and run it later". Because the original data is already gone (assuming no other transaction had a read lock on it), all the updated data not only has to be deleted but then the redo logs replayed to get back to the original state.

Thank you!

[jd]

The link is most helpful and I'm glad to say that I did see that Linux is indeed still supported and they do still offer a trial version for it. I'm downloading all 133.5 megabytes of it as I type, eternally grateful I'm not on my original 300 baud modem.

Oh man! I haven't laughed that hard in years!

[Anonymous Coward]

You keep using that term: sql standards compliant database engines. I do not think it means what you think it means.

"MySQL," "standards-compliant," and "SQL" do not belong in the same sentence unless the word "not" is also present.

Depends on the XML

[Anonymous Coward]

If you have document-oriented XML data then I agree whole-heartedly. However, data-oriented XML is often times a clean fit.

For example, most XML data on the web in the form of XHTML would not be suitable for a relational database because it's document-oriented. On the other hand, this kind of XML is very well suited to a relational structure because it is data-oriented.

In the end, XML can be a good, non-DB-specific serialization method for relational data.

Then again, I think there's something to be said for native XML storage types within a traditional SQL database. Something like:

select * from foo where xpath(xmldata, '//some/xpath[@pattern]');

where "xmldata" is a table column that holds XML documents or fragments. For database engines that can be easily extended to add custom types (such as PostgreSQL), you could have the storage model for the XML data type use the exact same model used in dedicated XML databases. In other words, it's not an either-or situation; use both at the same time. Best of both worlds if you will without diminishing either.

Re:OK - I'll bite - the 2 proc limit is significan

[Anonymous Coward]

I don't know about DB2, but I am the DBA of SAP R/3 installation with 110 users, Oracle 9207, running on TWO 600 MHz Power PC CPUs (7026-6h1). Performance is under 1 sec for most results. I don't think the limitation of 4 GB and 2 CPUs is significant.

Too late IBM!

[Anonymous Coward]

This is the second time IBM have released a free version of DB2, at that time for Linux. Limping along with a very buggy MySQL, I suggested to my boss we upgrade. Then, I kid you not, IBM changed their mind and and silently backpedaled. Thanks for making me look like a Goose, IBM.

And since the Great God of Databases, Oracle, now have a freebie version, why would I want to dance with yes-no-maybe IBM again? I did look at their so-called Personal DB. Ugh! What a mess. One of worst user interfaces I have ever seen. Want to create a database? Good luck. 40 Start Menu entries, with names like Information Center, Control Center, Information Catalog Center. It uses Java for all its GUI crap (when I see the Java coffee cup appear in the corner of a window, I sigh and go and make one, because it is so bloody slow!) And it trashed all my Visual Studio macros without warning.

So, no IBM. Go away. We don't need you. Oracle has a free version now, and while they also have slow Java and bloated menus at least at least I can contract for $150 an hour with Oracle. DB2. Who the hell uses DB2?

Sorry IBM, but you couldn't give it away.

No OS X version shows IBM's cont. lack of support

[Been on TV]

IBM's continued more or less total lack of software support for Mac OS X can be seen as a telltale sign of why Apple is in the middle of making the switch from IBM's PowerPC processors to rival Intel's new generation of processors. I have blogged before on the background of why Apple is choosing Intel processors over the, in my opinion, in many ways still superior PowerPC/POWER IBM processors.

My take on this is that when Apple's XServe with the G5 started popping up in supercomputer configurations around the world at a fraction of the cost of similar configurations from IBM, IBM could only see this as a serious threat to their midrange and Unix product lines and could not allow Apple to undercut their business. It is no coincidence that the XServe has not been upgraded with the dual-core PowerPC 970MP like the Power Macintosh workstations has been.

Likewise, IBM Software has not ever lifted a finger to port their software to Mac OS X Server and help Apple promote IBM's own processor architecture. No support for DB2, Tivoli, Tivoli Storage Manager (with the exception of a half-baked client) or any of the WebSphere modules to mention a few.

Who of the parties actually cut the cord last year is still unclear to me, but I don't think Apple any longer could live with IBM reserving their best processor performance for their own hardware and thereby limiting which markets Apple could enter. Intel, fortunately, does not have such split priorities, which is good for Apple.

It will be interesting to see if IBM Software will be more eager to support Mac OS X once the switch is over. Presently they can hardly blame it on software difficulties given they market the excellent IBM XL-C and XL-Fortran compilers for Apples G5 hardware. A recompile of the Linux version of DB2 Express-C for OS X should be a no-brainer for IBM as a technical exercise.

Open Source databases

[Austin Milbarge]

have come a long way. Personally, I'm not a huge fan of open source, but for some things like databases and web servers, you can't deny that it has left it's mark in a big way. Now, I'm not very familiar with DB2 or SQL Server, so I won't comment on their features or lack of, but I've been introducted to PostgreSQL a few years ago, been using it ever since and I just can't see why (at least for small to medium sized applications) it wouldn't work just as well as an SQL Server or DB2. Postgres is awesome. You can even get commercial support for Postgres. In fact, I just heard that Sun is now including it in it's Solaris 10 system. Thats good enough for me. I don't see why one would shell out $$ for a database anymore, except for support.

Req Help: C# / MSDE - SQL 2005 Express

[Anonymous Coward]

We have an app written in C# (Visual Studio.NET 2003). It uses MSDE 2000 as the database.
It is installed on all our users Windows XP (Pentuim 4, 2.5 GHz, 256 MB RAM, 40 GB HDD)
computers. We used Microsoft Data Application Blocks (not the newer Enterprise Library) as
the database abstraction layer. The app is sweet and the users love it.

Now, we reached the 2GB treshhold. We cannot import in any more data from our
legacy systems. Our suggestion was to delete transaction data that was more
than three years old. We did not get permission. shucks.. That means we should upgrade to
SQL 2005 Express which has a 4GB database limit.

Upgrading the database by moving all the tables, stored procs and udf's is no big deal.
An added bonus is an easier installation for our IT support team , and a free database
management tool (SQL Server Management Studio Express?).

The big question is : can we retain the same source code base ie. will DAAB be compatible
with SQL 2005 Express ? what are the alternative solutions? Thanks in advance.

Cheers from Nigeria,
Steve Faleiro
Programmer.

Never mind

[msobkow]

I thought about it for a while, and realized what construct a full join refers to. I've never needed one, but I think a three-way union would do it ( left+right, left+null, null+right ).

Re:Never mind

[zardo]

Yep, I use them for comparisons, for example, I had a list of orders a supplier sent me and a list I made myself, because I wanted to see how crappy the list the supplier gave me was, so on one side I had a list of orders they supplier gave me that shouldn't have been there, the ones we both got right, and then the ones I had but the supplier didn't. Not much practical use in my experience, more for analyzing stuff, but if you're shopping for the perfect database you look for stuff like that.

By the way, all you need to do is a left join and a right join, using "union" (and not "union all") would eliminate any duplicates. This is the preferred method because it is faster.

Re:Never mind

[msobkow]

I wasn't actually talking about full joins for two of those limbs.

left+right is an obvious equality join.

left+null is left with null values for the right columns, where left is not in right.

null+right is right with null values for the left columns, where right is not in left.

That gives the database simpler constructs to evaluate while preparing the result set than a pair of outer joins.

OTOH a decent optimizer should take either form and end up doing the same internal operations.