IBM Needs to Act Boldly in Database Field

By Charles Garry  |  Posted 2006-02-15

IBM Needs to Act Boldly in Database Field

So IBM finally showed some signs of life in the distributed space when it offered DB2 Express-C as a free download.

As I noted in a previous column, IBM desperately needed to shake up the status quo in the Unix, Windows and Linux database market. Bless them if they didnt finally make a move.

Now many will say that this move by IBM is not a big story when you consider that Microsoft (MSDE, SQL Server 2005 Express), Sybase (ASE express), Oracle (10g Express) already beat them to it, in some cases by several years.

By the way, how did every marketing person on the planet come up with the same name? Why did they all chose "express"? How about "free," "cheap" or "crippled"? Its probably because marketing and truth are concepts with only a tangential relationship at best. In any case, I believe this announcement is a very big deal.

While virtually every other commercial database vendor offers a free version today, all of them have some sort of severe limitation. Most are intended solely for developers with the thought that if you win their hearts and minds, you win the database war. All of this is true to a point.

Click here to read the details about IBMs introduction of a "free" version of DB2.

However, what makes the DB2 Express-C announcement so important is that it provides not only a development platform, but also a deployment platform. This in my opinion is a huge differentiator.

With a limitation of two processors or four cores, IBM has touched upon a growing sweet spot for database applications. Consider that a two-processor x86 server today is more powerful than an eight-way RISC server was in 2001.

This means that more and more database applications could easily run their workloads on a current-generation two-processor box. With no limitations on users and database size, DB2 Express-C could realistically fill the production needs of as many as 50 percent of current database workloads. That, my friends, makes this significant.

This move places additional pressure on the other commercial vendors to meet this new standard. Of course they will wait to see if this move pays dividends for IBM in terms of market share. Theres no sense giving it away if DB2 on Windows and Linux remains a distant third.

IBM is pulling out all the stops however. Their DB2 Express-C download site offers some good supporting materials to help people get started. They have done the leg work, getting partners to announce the usage and bundling of DB2 Express-C.

I think the timing is significant as well. Oracle has been struggling to find a licensing model that fits todays multicore world. They have tried 75 percent solutions and 25 percent solutions for certain processors.

Next Page: Adapting to the open-source world.

Adapting to Change

These moves have simply added complexity to licensing at a time when the market wants clarity. I believe the move by IBM is only the next step toward the inevitable in database licensing.

I maintain that commercial vendors will be forced to offer license-free versions of their software for servers of up to 4 sockets. This is the dividing line between broad market and specialty high-end needs.

This is a land grab for market share, and commercial vendors need to adapt to a world where open-source databases can address many average workload needs. What open source cannot address is the major vendor support value chain. Open source is also never likely to cater to the needs of the very high-end database applications in a world where users would prefer to standardize on a small number of database platforms.

With Oracle primed to move aggressively into the "open source as commercial enterprise" space via its acquisitions of Innobase and Sleepycat along with the rumors of the further acquisitions of JBoss and Zend, the time is now for the move toward support-only models.

Read more here about how other open-source vendors are reacting to the introduction of IBMs DB2 Express-C.

While I praise IBM for raising the bar with the two-processor (four-core) limitation, its not bold enough. First, it is not offering companies a support-only option. To get that you still must buy a DB2 Express license. It is a move that defies reason.

It only confuses the market when IBM offers both a DB2 Express licensed version and a DB2 Express-C free version that are virtually identical, with the exception that you cannot buy a support offering or get the nifty HADR (High Availability Disaster Recovery) option. Why not gain market share and be able to upsell support packages and options such as HADR?

Click here to read about IBMs "Viper" DB2 Information Virtualization Server.

Dual licensing might have made sense for MySQL because of the very nature of the GPL. But there is no such concern for DB2 Express-C. Without eliminating the distinction and offering a fee-based support option, DB2 Express-C comes off more as a gimmick, such as those offered by the other vendors.

The free community support is great for developers, but IBMs core audience (Global 5000) will want professional, guaranteed support when they move to production. This is what has held back wider open-source adoption in the enterprise space. Who wouldnt want IBM standing behind a product? Not to mention the margins IBM could make on support and the wider adoption this would drive.

So Im cheering for IBM today, but I warn them its not enough. If you want to lead you have to be bold and look at the bigger long-term picture. I dont believe that future will have room for traditional software licensing models. Why not lead the change?

Charles Garry is an independent industry analyst based in Simsbury, Conn. He is a former vice president with META Groups Technology Research Services. He can be reached at

Check out eWEEK.coms for the latest database news, reviews and analysis.

Rocket Fuel