Data Domain sticks neck out on deduping
Dopey notions from speed freaks?
Solid state storage represents a great big opportunity to Slootman. "We're very, very interested in that. It's not electro-mechanical and it's blazing fast, but it's got economic problems. Deduped disk is destroying the tape market. Deduped SSD will affect the Fibre Channel disk market, giving SSDs the economics for the mainstream market place. We think dedupe will become a huge enabler for SSDs."
Fast transaction data deduping was specifically linked by him to virtualised servers and their virtual machine images where there is a lot of data redundancy.
When might this happen, this combination of deduping and solid state storage in Data Domain arrays? By the end of 2010? "It's not out of the realm of possibility."
He'll have half an eye on NetApp which is going to introduce SSD technology into its arrays, and which reckons WAFL technology inside its ONTAP array software is already well suited to writing data to flash memory. NetApp also has its ASIS dedupe technology shipping with every ONTAP array it builds and has started suggesting it's OK to use it for light transaction data use. Here are building blocks NetApp could use for a Data Domain catch-up effort.
Data Domain builds storage arrays and if the company manages to position them as general storage arrays, not just for backup and archive use, then they will be judged and compared as storage arrays against competing product from Dell, EMC, HDS, HP, NetApp, Sun and others. So Slootman has to ensure Data Domain builds out its software environment so the company's products can compete in the general storage array market. While he's doing that, though, the company is not going to stop pressing the gas pedal to the floor and accelerating its arrays' storage operations as fast as possible.
While talking the general storage array talk, Data Domain will also be walking the accelerated dedupe array walk, with its products positioned now as fast and affordable replacements for tape libraries and, potentially in the future, for fast and affordable deduped, HDD and SSD-featured arrays for transaction data storage at block level or, we guess, file-level. The company has to use the profits it's generating now to build products encapsulating a strategy that will enable this one-trick pony to broaden its offerings and widen its trick portfolio. Speed is the key - and that means an opportunity for its competitors.
If they can add similar 8-core Xeon controllers and rewrite their array software to match Data Domain's speed, then Slootman's edge withers away. It really is a race dependent on speed. Can Data Domain use its current speed advantage to earn the revenues needed to bulk out its offer, broaden its product range, and get a solid place in the market, before Slootman's competitors add the processing power they need to their deduping arrays and stop Data Domain in its tracks?
Gentlemen, rev your engines... ®