“The creators of the open-source databases MySQL and PostgreSQL are trying to push them further into the enterprise with new features aimed at better support for transactions, database recovery and replication.” Read the news at eWeek. On the other hand, VA-Software said that they had to switch SourceForge to IBM’s DB2, because their database grows very fast (70 new projects and 700 new users daily) and while they did a move from mySQL to PostgreSQL a year ago (mySQL is faster for smaller dbs, while PostgreSQL scales better), now they have to move again to an even more powerful database, which happens to be proprierty.
Open-Source Databases Hike Enterprise Appeal
2002-08-20 Databases 13 Comments
Their are some areas where proprietary databases have an advantage over the Open Source ones. For the time being that is. PostgreSQL has made remarkable progress and I fully expect it to match Oracle and DB2 in a couple of years. MySQL is nice too – for small projects. Today the best tool for Source Forge may be DB2. Use the Right Tool for the job. Though, might VA Systems be trying to court a buyout by IBM?
yay! go DB2!! i love working with db2. hell i just love as/400’s heheheh
PostgreSQL has made remarkable progress and I fully expect it to match Oracle and DB2 in a couple of years
Omer, I persnoally cant see Postgress matching DB2 or Oracle. I think it will struggle to ever match SQLServer
Trackrecord is important as well, and while I’d trust my datas to DB2 or Oracle anyday (major banking applications have been using them for ages) Postgresql still has to prove its reliability (especially since peoples in the past had pretty serious data corruption). This is especially difficult since so few peoples use Postgresql, which means few people can witness it’s strength in production environement and testify about it.
As much as I find Postgresql a useful DBMS in some circumstances it just can’t be compared with Oracle. You can and people do bet their businesses on Oracle, you just don’t want to do this with Postgresql, it just isn’t there yet.
This was the press release concurrent to the db2 migration
SAN FRANCISCO–(BUSINESS WIRE)–Aug. 13, 2002–VA Software Corporation (Nasdaq: LNUX – News) today announced a significant commercial agreement with IBM (NYSE:IBM – News) focused on the joint marketing and sales of the next generation of SourceForge(TM) Enterprise Edition, which will offer full support for IBM’s DB2 database. SourceForge is the leading collaborative software development environment used by more than 460,000 users worldwide.
As part of the agreement, IBM’s direct and indirect sales channels will engage with VA Software in joint sales activities. Together, VA Software and IBM will work to solve the complex application development challenges faced by enterprise customers in distributed environments
The terms of the agreement include VA Software’s and IBM’s ongoing integration of SourceForge with a suite of IBM infrastructure, application development and system management tools, including DB2 database software; WebSphere Application Server; WebSphere Studio Application Developer; and Tivoli management software. In addition, SourceForge will also be optimized to run on IBM eServer xSeries Linux servers. SourceForge Enterprise Edition for DB2 is expected to become available before the end of the year.
read the rest here: http://biz.yahoo.com/bw/020813/132117_1.html
They don’t need to move to more scalable software, they just need to do some routine maintenance and cleanup of the thousands of projects that haven’t been touched since 1997. The great majority of the stuff there is at Revision 0.00.00.01. Right now, SourceForge is the coding equivalent of Yahoo’s free email accounts. Everybody’s got two or three projects on the go, because they’re free to start up. But nobody, including the lone developer, ever looks at them. We don’t have enough endurance to complete a project on our own. We have lots of inspiration and good intentions the first little while, but that dies out quickly due to overwork and lack of interest.
SourceForge doesn’t need new projects as much as it needs to facilitate the completion of projects already started; get some organization and teamwork set up to improve the quality and completeness of their projects. Have an organized bulletin board where coders can see what kind of projects are being started (by category) and how they can help out. There are too many lone coders, and not enough team players. There is too much duplication of projects and code. Why make 20 poor text editors, web browsers, or Amiga emulators, when we could work together and pool our resources and make one Awesome app? Most of SourceForge’s stuff is unusable, and has been for years, and will be until the SourceForge system is changed to encourage teamwork and code reuse.
And THE most important part of a project is finding a good successor to take the reins from you when you are no longer interested in the project!
Just my two cents, which I think is greatly undervalued.
With IBM moving to Linux world it’s just like free advertisement for them.
They announced that it will run on IBM linux boxes too.
VA did a smart move – it’s a great way to keep SourceForge alive. IBM did smart move too – who would want Oracle on Linux if DB2 is so good that even SourceForge is running on it. As I don’t care too much for database war I’m quite happy with SourceForge landed to safe harbor in these horrible economic times.
…is “World Tracer”, running on Oracle in Atlanta. Almost all airlines use it to track stray baggage.
The DB we have saved our Windows-based project with is Firebird / Interbase. I’d love to know how it compares with PostreSQL. We found MySQL much slower, but we were running it with Borlands dbExpress components, whereas Firebird was being read through native InterbaseExpress components. We had been using Access / ADO before, and not surprisingly it was slow and had the odd data corruption.
I love these articles like the one at eWeek that put PostgreSQL in the same place as MySQL. Oh, and transactions are a big deal? They should be a given. PostgreSQL has always had transactions.
I agree that there are some scaleability and feature issues that PostgreSQL doesn’t handle quite as well as some of the larger DBMS’s. But, PostgreSQL does handle logical manipulation of data, and data constraints quite well (in fact, it has a few neat extras that the Big 3). In other words, the fundamentals are all there, which cannot be said of MySQL. The only areas where PostgreSQL really needs to catch up are scaleability, schema support, and replication. These sound like big deals, but for the most part they are not logical problems, just quantitative ones, and PostgreSQL is already doing well in each area.
The things I love PostgreSQL for are:
1. Low footprint. PostgreSQL definitely gets you the most bang for the bytes of any DBMS. 9.5 MB for the full source download. The postgres daemon usually uses less memory than MySQL on my servers.
2. It is the only real RDBMS which runs natively on FreeBSD. I consider this a serious advantage, because FreeBSD is one of the most rock-stable OS’s on the planet. I want my database on a stable platform. That’s more important than performance. (PostgreSQL’s C source is so clean that it compiles and runs on just about any Unix)
3. Great logical manipulation features. Performance isn’t everything. While other DBMSs compete for TPC benchmarks, PostgreSQL allows me to do some amazing things with the actual data itself. User-defined datatypes, query rewrite RULEs, Geometric (3D) data manipulation, sophisticated date and time functions (a datatype called INTERVAL(?)), and much more. Play with it for awhile and you will be hooked. In many ways it is the most advanced DBMS on the planet, especially for science/research projects.
Any idea on DB2’s pricing on Linux ?
Would this Linux version run also on any of the *BSDs ?
If anyone knows,thanks for your advice,
We use Oracle for a large and very complex application running on clustered and concurrently working Sun servers + large number of fiberchannel disk enclosures etc. I was amazed at the robustness of the solution: each node performs transactions servicing requests in a round robin fashion, yet the database never corrupts (and these are complex, “long” transactions). You can kill any of the nodes, any of the disk enclosures, any of the disks or even any of the fiberchannel switches, and the system will continue working, and the database won’t be corrupted.
Some of the features of Oracle used in this solution have been made possible by good cooperation between Sun and Oracle developers.
As for the application itself, I can only disclose that it’s for large mobile operators.