Intelligent Enterprise

I heard about an interesting presentation planned at the EMC World conference in May. According to this talk, EMC will introduce another entry into the warehouse appliance market via a partnership, this time with ParAccel:

"EMC and ParAccel have jointly engineered and developed a highly scalable and performant analytic appliance. This solution is built on EMC CLARiiON midrange CX-3 UltraScale networked storage and ParAccel's analytic columnar data store. Customers can deploy the EMC/ParAccel analytic appliance by simply extending their existing EMC footprint on enterprise ready storage while leveraging EMC's proven solutions."

I think this is notable for a few reasons. On the competitive front, it's going to put pressure on the hardware-based appliance vendors — companies like Netezza and Dataupia. The concept they sell is inexpensive scaling and capacity for large data volumes. If a storage vendor like EMC is capable of doing similar work approaching a similar price-point, it's going to be hard to stay competitive as an appliance business. Particularly if it's as easy as EMC implies to "extend their existing EMC footprint."

Hardware-based appliance vendors face potentially higher costs due to their small scale and the engineering required for custom hardware. More important, it's hard for them to keep pace with ongoing performance improvements in commodity hardware. The one cost advantage these vendors do have is somewhat lower support costs, and they don't need to worry about porting and certification for multiple hardware platforms and operating systems, as software-only vendors do.

DATAllegro moved away from a hardware-based solution last year, partnering with EMC and Dell. As detailed in this article surveying the market, many other competitors — ParAccel, Vertica, Greenplum, InfoBright — pitch horizontal scaling on commodity hardware.

EMC's new appliance puts a kink in the DATAllegro-EMC partnership. If it's this easy for hardware vendors to mix in other alternative data warehouse databases, then it's going to be hard for DATAllegro or any other software vendor to do exclusive hardware deals and stay competitive with peers.

My bet is that EMC will position the DATAllegro appliance for bigger data volumes and ParAccel for high-speed query at the next step down in scale. Horizontal scaling in row-store databases has been with us for years. We haven't seen large-scale implementations of columnar databases yet, so the high-end volume story is yet to be determined.