Today, Oracle announced that it is buying SAN array vendor Pillar Data. Pillar has, since its inception, been funded by Oracle CEO Larry Ellison’s Tako Ventures, so the whole transaction comes as a surprise to no one who follows the storage industry. Despite the Ellison connection and its innovative architecture, Pillar managed to sell only 1,500 systems to some 600 customers. It will have to do better than that if Pillar CEO Mike Workman is ever going to be able to cash his stock options in, as Oracle isn’t paying cash or stock but offering a buyout based on how Pillar does in three years.
Martin Glassborow, who blogs as Storagebod, asks and answers the $64,000 question: "Does this have any real meaning for the rest of the enterprise storage world? I suspect not, things are moving on quickly and, Pillar will simply be consigned to the annals of history as a rich man’s play thing. It's probably more proof that Oracle really has little clue about the hardware side of things and should have stuck to software."
I think many, including the few diehard Sun loyalists who were left when Oracle bought that once high-flying hardware vendor, agree with Mr. Bod that great software companies probably shouldn't more than dabble in hardware. While Microsoft’s CP/M card for the Apple ][ did bring IBM to Bill Gates' door, Microsoft has made a lot more money on Exchange and SQL Server, let alone Windows and Office, than on Xboxes and mice.
For Oracle, Pillar's Bricks and Slammers will replace the SAN arrays Sun OEMed from Hitachi Data Systems and Engenio. Except for the high end, where customers will probably continue to buy HDS Universal Storage Platforms, Pillar's systems are probably an improvement over the unexciting Engenio systems.
The bigger story, I think, is that this represents the end of the RAID 2.0 era. During the past few years, innovative startups--including 3Par, Compellent, Xiotech and Pillar--have advanced the disk array from drive-based RAID to more flexible "chunklet" technology that integrates data protection and data placement across all available drives. Xiotech and Pillar extended this further into a hybrid scale-out architecture by integrating the RADC (Redundant Array of Data Chunklets) controllers into the drive shelves.
With Compellent and 3Par acquired, the RAID 2.0 story seems about over. Future innovation for the midmarket to enterprise level will come from solid state for speed and cloudy technologies like dispersal coding for long-term storage at reasonable prices. RAID 2.0 was the last hurrah of the high-speed spinning disk.
Since Pillar CEO Mike Workman is a serious amateur pyrotechnician, I hope he will put on a good fireworks show this Independence Day in celebration.
