Linux will become the dominant OS for Oracle databases by 2007, displacing Solaris, according to a new survey from the Independent Oracle Users Group (IOUG). But just by a smidge -- 44 percent will run on Linux by then, versus 43 percent on Solaris. Of the users who responded to the survey, 34 percent planned database migrations in the coming year. "This forecast marks a definitive tipping point for Linux," trumpeted the press release.

OK, but of the 14,500 Oracle DBAs the survey was e-mailed to, the IOUG received only 812 "usable" responses, or a whopping 5.6 percent. So what kind of systemic bias might be baked into this result? Maybe only the Linux enthusiasts responded? Maybe the Solaris folks had no systems issues that day and went to the beach?

Although not a great sign for Sun -- especially if, as rumored, Oracle adds a Linux distributor to its stack -- the real story here is how fast commoditization is happening. One by one, the great proprietary software franchises are coming under intense pressure from low-cost alternatives. Red Hat acquiring JBoss will surely speed this process. Who knows how fast a big franchise can implode when a tipping point is reached? It will be interesting to see.

The Google Enterprise Effect, Take II

Meanwhile, Google is making enterprise IT news again. You've seen how that company has changed the world in our personal lives -- how did we ever do celebrity photo searches before Google, anyway? Now Google's wading deeper into enterprise IT, pushing past its initial appliance offering.

Essentially, Google has expanded its abilities to look into enterprise applications and pull out data based on keyword searches. According to a recent AMR Research report, "The Google Push (http://www.amrresearch.com/Content/View.asp?pmillid=19384)", the search giant may be making a play to become the presentation layer for critical enterprise data, potentially disintermediating enterprise software vendors.

The new version of the Google Search Appliance, notes AMR, is a big improvement -- extending its reach into various structured and unstructured data sources, integrating more cohesively with enterprise security infrastructure, and improving performance. And the Googlers aren't expected to stop there.

"Google recognizes the importance of the end-user point of view, both as the source of its own success in the consumer market and the obstacle to success of many enterprise systems," writes AMR's Jim Murphy , the report's author. In particular, AMR sees opportunities to add value in making business intelligence, CRM, and ERP deployments more accessible and usable.

But to do so, Google will need strong enterprise partners to lend it enterprise credibility, AMR argues, including SIs and software vendors. And why would these partners want to let Google take the presentation layer away from them, I wonder?

Google must also change perceptions about security and privacy. I recently lunched with a group of large-enterprise CIOs when the topic of Google's new enterprise capabilities came up. "I'd love to have better enterprise search," said one CIO, "but I'm not gonna let anybody store my data outside my firewall to do it."

Personally, I'd love to see Google make some headway in the enterprise. I think it would have the same effect it's having on Yahoo and Microsoft: better services, lower prices, and better support.