Open Source Database Consortium
The database market is experiencing some changes. All major database players now have either a no-cost database, or even an Open Source database. A month ago, that was still not the case.
Personally, I think it’s a good thing when there are new, lower-cost database alternatives offered to customers. However, Free and Open Source Software (FOSS) is about much more than “cheap”. Market research (such as Evans Data) so far indicates that “lite” versions haven’t gained much traction. I believe the reason is that most users are looking for a full-featured database product, not a handicapped trial version.
As a footnote related to the power of OS DBs, Business Week reports about the Evans Data survey as follows, 21 October 2005:
Mainstream database software companies that have been watching open-source upstart MySQL in their rearview mirrors must be having some uncomfortable thoughts right about now. According to a user survey released Oct. 20 by market researcher Evans Data Corp., open-source database deployments in corporations are up 20% in the past six months, and use of MySQL is up 25%. Forty-four percent of corporate software developers surveyed said they use MySQL software. In the coming months, MySQL’s growth rate could hockeystick.
The Open Source Database Conference in Frankfurt earlier this week provided me with an excellent opportunity to interact with other providers of FOSS databases. More precisely, with the help of Georg Richter and Zak Greant, I had invited a number of FOSS DB developers to a dinner last Tuesday 8 November 2005 opposite the NH Mörfelden hotel in Frankfurt (where the OS DB conference was being arranged) to test the hypothesis whether the ground is fertile to establish an Open Source Database Consortium as a (formal or informal) vehicle of exchanging information between FOSS DB providers, and as a body for the OS DB industry.
So we met, in alphabetical order (sorry if I missed someone; just tell me and I’ll correct it):
- Brian Aker / MySQL
- Kaj Arnö / MySQL
- Patrik Backman / MySQL (indirectly representing Jörg Hoffmeister / MaxDB, who had to leave just before)
- Josh Berkus / PostgreSQL
- Marcus Boerger / PHP
- Gregory Burd / Sleepycat (BDB)
- Daniel John Debrunner / Derby
- Peter Eisentraut / PostgreSQL
- Zak Greant / SQLite (as proxy for Richard Hipp)
- Lenz Grimmer / MySQL
- Øystein Grøvlen / Sun
- Jutta Horstmann
- Anders Karlsson / MySQL
- Holger Klemt / Firebird (with Ann Harrison and Paul Beach invited but not able to come)
- Arjen Lentz / MySQL
- Mauricio Longo / Firebird
- Georg Richter / MySQL
- Gavin Sherry / PostgreSQL
- Lukas Kahwe Smith / PHP
- Fred Toussi / HSQLDB
The discussion was lively and any open animosity between the representers of competing OS DBs, if there ever was any, disappeared after the first sip of wine, Weissbier, or whatever we had. A lot of technology information on cool projects, such as PostgreSQL’s Slony-II replication, flowed as freely as Open Source information flows.
In an at least partially successful attempt to impose something vaguely resembling an agenda on the group heading almost 20 people, I encouraged smaller groups auto-forming in the various parts of the table to structure their OS DB Consortium thoughts into three phases:
- Brainstorming: What are the things we could do together? What are our common interests? From what type of cooperation amongst us could our users benefit?
- Evaluate: Which ideas are worth considering?
- Low-hanging fruit: Which ideas can we implement easily?
As an additional item we discussed our legal status: Should we have a web site? A legal organisation, with by-laws?
1. The brainstormed ideas were plentiful:
- Joint benchmarks to make it easier to compare OS DBs
- An external advocacy mailing list
- A private mailing list for emergencies, security issues (such as zlib)
- Joint development: standards (such as PDBC) and APIs, research, conformance test suite, performance test suite
- Joint marketing: announcements, advocacy of FOSS licenses over closed source licenses, approaches to universities
- Joint PlanetXxx.org blog aggregation
- Joint legal research and statements on SWPATs, other IP issues, licenses
2. When evaluating ideas, we noticed that:
- we have different stances on SWPATs, ranging from the clear, strong and open support (and even financing) of NoSoftwarePatents.com, to concerns of the wisdom of a clear NoSoftwarePatents.com support when being financed by some of the world’s largest SWPAT holders; although it was hard to find SWPAT proponents around the table, we could deduce that NoSWPAT advocacy is not a common denominator
- although it may be desirable, it is hard to define a fair common benchmark (but the one c’t is using, the Dell DVD shop, comes close)
3. The low-hanging fruits were suggested to be:
- The establishment of a common website, OSDBConsortium.org or something (thanks Gavin for already implementing this!)
- An internal mailing list for our own discussions on our future ideas
4. As for the legal organisation,
- we could have individual members, or corporate members, or both
- we could see the Apache Software Foundation as a model
- we could leave the legal organisation until later
If you were present, or you’re just interested in being in this consortium, leave a comment in this blog article!
November 12th, 2005 at 22:56
On a personal level it seemed like everybody got together quite well. However we quickly noticed how difficult it was to get any sort of decision done. For the most part this seemed to be due to the fact that people needed to get the feedback from their relevant projects/employer.
It also seemed that it is not quite as simple as the enemies of my enemies are my friends on a project level. As such the cooperation or consortium may end up being just a number of individuals from various projects. This would of course mean that we would not get the same kind of punch together, but if we can a nice enough mix together it could still gain alot of credibility with users both in the OSS and in the commercial world.
November 13th, 2005 at 17:12
Denis and I were sorry we couldn’t attend this first meeting. There is certainly work we can all do together, even if it is simply to promote the enterprise-readiness of open source (which many people know already). Keep us in the loop and count on us for support.
November 14th, 2005 at 1:17
Kaj,
First off, even though Gavin has already pegged down the domain, I’d like to suggest that “Open Source Database Association” is actually probably a better name, since Consortium implies a cohesive incorporated body and specific joint ventures (at least in American English) whereas Association is less binding (and thus more applicable). However, it may be a bit late for that comment.
Second, Ingres was not present due to a miscommunication about the event.
Can someone contact them?
Regarding Legal Organizations: at the meeting, I believe that there was a lack of consensus that a formal legal structure for an OSDBC was even required, let alone any consensus on what form that organization should take. I commented there, as I will here, that the ASF is an organization designed to produce software which is completely different from our needs. If you’re looking for a structural example, the OSDL or Web2.0 group would be the closest. My opinion, though, is that we shouldn’t bother looking at legal structures unless/until our group decides to hold a conference or something.
Regarding Benchmarks: I think we had a completely different discussion about this on Monday which unfortunately didn’t become part of the record on Tuesday. So here it is: all DB developers I discussed this with can agree on the value of a portable suite of database benchmarks, possibly starting with the various TPC ports done by OSDL and Jan Wieck. That, for that matter, we have a collective interest in influencing the TPC to become more friendly to organizations that can’t afford to spend $200,000 on a test, and could work with TPC in making their benchmarks more OSS-DB friendly. I didn’t get any argument on this one, so I think it’s actually something that everyone can agree on.
More later, on the mailing list/wiki, I suppose …
–Josh Berkus
PostgreSQL Project member
November 18th, 2005 at 0:17
Hi
Ann and Paul where running to the Firebird World Conference in Prague, but a Firebird Foundation representative was there: Mauricio Longo, the first from the left in the Photo.
Mauricio already talked about this meeting in Prague.
Artur Anjos
Firebird Foundation Member
November 18th, 2005 at 19:35
Note that lukes is on other side of camera (he take the photo, don’t search him)
Newbies request : all Open Source Database was present ?
November 18th, 2005 at 22:43
Hi, I think the next revolution in database will be the introduction of AJAX web applications that will access php mysql databases on the fly as the viewers actions trigger events.
January 27th, 2006 at 19:47
Hi Guys,
I just spoke to Kaj and I now wanted to express db4objects’ intent to join the party — unless you feel this should be a relational-only club.
db4o (http://www.db4o.com) is the open source object database which has entered the open source stage in November 2004 to focus on embeddable persistence in Java and .NET environments.
Use cases are devices like Boeing’s aircrafts, BMW’s cars, and Seagate’s consumer electronic devices, but also mobile enterprise applications, where business objects get bi-directionally replicated from the RDBMS servers via Hibernate to be persisted in distributed, partially connected environments with low memory footprint (400K) and zero-admin capabilities. In short: we feel we are a complement to (most) relational databases, not a substitute, so there should be grounds for cooperation.
We use the dual license model like MySQL and Berkeley DB (anybody invited them?) Our forums are here: http://forums.db4o.com/forums/
Comments are appreciated.
Christof (christof at db4o dot com)
February 1st, 2007 at 15:41
[...] A while ago, people from the major Open Source database systems have met to form the Open Source Database Consortium - that was in October 2005. [...]