Why Ingres Is Wrong on Oracle and Right on Oracle

By Lisa Vaas  |  Posted 2006-04-07

Why Ingres Is Wrong on Oracle and Right on Oracle

At LinuxWorld earlier the week of April 3, I hooked up with Dave Dargo, once vice president of Oracles Linux program office and now chief technology officer and senior vice president of strategy for Ingres.

He shared some interesting thoughts on new markets for Ingres, but before I go there, I need to pop a silly Oracle-related marketing bubble put up by Ingres earlier this week when it announced a partnership with Cognizant.

Background: Ingres is, like the PostgreSQL-based EnterpriseDB, an open-source database player hell-bent on prying away Oracle customers.

Sony Online Entertainment ditched Oracle in favor of EnterpriseDB. Click here to read Lisa Vaas thoughts on what it tells us.

Ingres announced on April 4 a partnership with IT systems integrator Cognizant to support the deployment of its database to Cognizants customers. This is all good: It represents a new market channel for Ingres, given that Cognizant has agreed to develop an Ingres database services practice as part of the partnership.

But heres the silliness: In his announcement of the partnership, Tom Berquist, Ingres CFO, claimed that a third of Oracles installed base is still running Oracle 7, and less than 10 percent have upgraded to 10g because of cost and complexity.

Dont swallow those numbers whole. The IOUG (Independent Oracle Users Group) just came out with a report that found that 84 percent of respondents are running Oracle 9i. A whopping 49 percent are running 10g, and virtually all respondents intend to update to 10g in the coming year.

I can believe in both these seemingly disparate numbers, because so many shops run multiple versions at one time.

The IOUG didnt list any numbers for how many respondents are running Oracle 7, but it wouldnt surprise me if 33 percent are. Along with Oracle 8, and Oracle 9i, and 10g, to boot.

And as far as the less than 10 percent having upgraded to 10g? That could well reflect the slow and steady pace of migration. It doesnt necessarily mean that 90 percent are frightened of 10g, as Berquist states, but could well indicate that a majority havent specifically "updated" yet, per se, because theyre still testing 10g. Thats more likely, given that 49 percent of shops are running 10g, according to the IOUGs numbers.

Click here to read more on the IOUGs report, which finds that Linux is moving to dominate Oracle shops.

Not that Im convinced the IOUG is unbiased, mind you. Its survey, called the IOUG Survey on Technologies for 2006 and Beyond, was sponsored by Symantec and produced by Unisphere Media, which publishes Database Trends & Applications and 5 Minute Briefing: Oracle e-mail newsletter.

Yeah, the IOUG is independent, but, well, theyre an Oracle user group. One could suspect they are Kool-Aid drinkers.

Next Page: Competition.


Ingres doesnt see MySQL as a competitor, and rightly so. Even though it gets more media buzz as an open-source player, Ingres offers a product with a longer history of enterprise deployment, so its definitely set up as a more serious alternative to Oracle technology.

They may both be open source, but thats about all the basis for comparison there is between MySQL and Ingres.

That pertains particularly given that theres no longer any uncertainty around MySQLs future vis-à-vis storage engines—at least, there shouldnt be, given that MySQL is going to roll out its own storage engine later this month at its user conference and has signed a multi-year contract with Oracle to keep using the Innodb storage engine, with the former license terms unchanged.

In a nutshell, I dont think Ingres has to scare Oracle or MySQL into its fold, and I dont think it has much grounds to do so for either database.

What it does have is some good ideas for new database markets. For one, Dargo, like all of Ingres top brass, is pitching Ingres as the Linux of open-source databases: Nobody took it seriously years ago, but look at Linux now.

The other market Ingres is aiming itself at is even more lucrative, Dargo says. It gets into the integration of maintenance. Whereas companies like Microsoft and Oracle are selling fairly complete stacks, the benefits arent getting to the customer, Dargo says.

"Customers still have to apply fixes to components," he told me. "Thats incredibly time-consuming for enterprises to go through, and theyre [looking] for simplified maintenance."

Because Ingres is playing in the open-source space, it can go out and integrate with other pieces of the stack, such as Linux, to make its own stack.

So whereas Microsoft is talking about extending the operating system to include the database, Ingres is talking about extending the database to include the operating system.

"I strip Linux to the basic functionality I need, and I can now run the Ingres database on bare metal," he said.

"Whats more interesting, virtualization comes around, and what Im delivering is a single entity called Ingres database that includes the operating system, and the customer doesnt have to apply maintenance to a separate operating system."

With virtualization and other technologies that affect implementation, its interesting to any company that has to rely on Microsoft or Oracle technology for a vendors product to be successful.

For an e-mail server competing against Microsoft Exchange, for example, Ingres could come out with an appliance that extends the database to include an e-mail server.

Thus, the appliance vendor can go to customers and sell an integrated software appliance thats both e-mail server and database, managed and maintained as a single piece.

The potential audience is any software vendor that has a solution that runs on Microsoft or Oracle.

This isnt pie in the sky. Dargo told me there are both customers and partners interested in this. A lab prototype is up and running in the lab, although Dargo didnt give me specifics (the e-mail appliance is just a for-instance).

Customer prototypes are on track for this summer, and by the end of year Ingres is planning to have a production Ingres software appliance. Dargo says that by next year well start seeing partners with software appliances.

I dont think Ingres "10g is too complicated and scary for a simple caveman like me" approach is going to get the company much business, but this I could see being innovative in a way that, according to Dargos thinking, closed-source vendors like Oracle just dont have the luxury to delve into.

By the way, Dargos blogging now, so go check out his thinking about the imminent demise of the closed-source business model ("The Broken Covenant of Software Licenses") and let me know what you think.

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

Rocket Fuel