Illuminata analyst Jonathan Eunice said the systems mark a new approach for IBM. "The whole model for how IBM goes to market is starting to change, and it will change pretty rapidly, with the idea that they are going to advantage their own software stack as opposed to having a server unit that's completely neutral for all of the ISVs [independent software vendors]," he said.

The systems follow the September launch of Oracle's Exadata Database Machine, which combines Oracle software with Sun hardware and storage technologies in a preconfigured system for data warehousing and online transaction processing. IBM's message on Wednesday echoed Oracle's when it launched Exadata -- that the systems are "integrated at every level -- from microprocessors to hardware and software."

Both vendors say configuring and tuning the systems in-house allows them to provide better performance and faster time to deployment. That may be so, but there are also trade-offs, analysts said.

"Customers are going to have fewer choices; they're not going to have the mix-and-match capability that they once did," Eunice said. "But the good news is, IBM isn't taking you back to the world of the '60s and '70s. You're not forced to buy the whole stack. If you want to buy an IBM analytics system and plug an Oracle OLAP cube in the middle, you can still do that. But for most customers it's not economically practical."

Dan Olds, principal analyst at Gabriel Consulting Group, agreed. "You don't have the opportunity with these systems to do the exhaustive shopping for best-of-breed applications that completely meet your needs," he said. But many IT shops are strapped for cash and can't afford the extensive service engagements required to put these high-performance systems together in-house. "If these products get you to 80 or 90 percent of where you want to be, then that's a pretty good trade-off," he said.