HOUSTON -- Carly Fiorina, chairman and CEO of Hewlett Packard Co., used her keynote speech at the company's StorageWorks Conference in Houston, Texas, Tuesday to outline HP's storage vision and stress its commitment to the industry.
Fiorina wasn't scheduled to be on the bill at StorageWorks, but after the company's
"I am here today in particular because some of you may be wondering about our commitment to storage ... our execution in this business has not been all we would have liked recently ... but our commitment is unwavering and we are in this business for the long haul," Fiorina said.
With that, she went on to describe HP's grid strategy, which is a loosely defined as utility-based architecture that uses "smart cells" for designing storage systems. Each smart cell is a self-contained system with its own storage, processor, search engine, database, indexing capability and management layer that can be connected to an infinite number of other smart cells, all in communication with and supporting each other. Provisioning in this smart cell world goes away, so when a new cell is added -- which can be given different personalities like file serving, block serving or archiving -- the other cells assimilate the new one and can automatically read its needs. Think of the Borg in Star Trek, and you're on the right track.
HP didn't pioneer this concept of cell-based arrays. Two startups that also offer this technology include 3PARdata Inc. and Isilon Systems Inc.Grid roadmap
HP's product roadmap for the grid strategy is as futuristic as the concept, according to analysts at the conference. Right now, it includes two recently introduced products: Reference Information Storage System (RISS), HP's archiving appliance and Scalable File Share (SFS), a global file system. A grid-based array is expected in the 2006-2007 timeframe, along with downloadable smart-cell functionality, seamless repository virtualization, and other whiz-bang sounding stuff.
Analysts said explaining this complicated message to users will not be easy. "None of it makes much sense today," said Mike Karp, senior analyst with Enterprise Management Associates. Stephanie Balaouras, senior analyst with the Yankee Group, reiterated the same sentiment, with the exception that if HP stays on message, users will eventually get it.
Chris Van Vliet, network systems engineer at Louisville Gas and Electric Co., doesn't see how the grid strategy solves his problems today. "It's a bit blurry ... HP has its hands much more than just storage," he said. Louisville Gas is looking at the RISS system and EMC's Centera appliance for its archive data.
Likewise, Jim Razmus, manager of database systems at Herman Miller, is skeptical of how HP's grid plans fit into his current storage environment. "It's very interesting, but it has no value for us right now," he said. Herman Miller is an EMC customer with 25 TB of raw storage on Symmetrix arrays. The company is about to reach the end of its lease and is exploring other options. Razmus is, however, interested in the XP 12000, the new high-end array HP resells from Hitachi Data Systems. "The continuous access software is pretty compelling, too," he added.
Greg DeYampert, manager of technical services at Healthcare Partners Medical Group, said he is more interested in where HP is taking the EVA product line, not what it's planning five years from now. "I need to make technology decisions today. Money is way too expensive nowadays, and I can't afford to make a mistake," he said. Healthcare Partners has 4 TB on HP storage right now and will need to buy an additional 8 TB to 12 TB over the next two years.
HP's answer to the future of the EVA is that it will be around for a long time to come with increased speeds, feeds and functionality in the works. But expect a smart cell head, which will have much of the intelligence of the grid array but without the disk, to creep in front of the EVA and start sucking its guts out as soon as 2005 -- an approach that is clearly not for the fainthearted.