EMC Invista Mark II
In May 2005 EMC announced Invista, their groundbreaking Storage Virtualization product. The only ground it appeared to break was it's own grave.
Then EMC buys YottaYotta; and the only ground YottaYotta broke was its own burial spot too. At least they had that in common.
So after a bit of tomb-robbing, five years later EMC is introducing its second attempt at Storage Virtualization. They present this under the seemingly cooler term of Federated Storage to try and make sure everyone forgets their first failure (who's calling them out on that piece of marketing chicanery?). Talk about dressing up the coffin.
And the irony of their announcement is the hypocritical focus on being able to virtualize IBM SVC. Don't forget for years EMC has waxed lyrical about how ridiculous it is to virtualize a Symmetrix or Clariion under IBM SVC. Now they are hyping Virtualization of the Virtualizer. Soon will have some other pseudo-clever company saying that they can Virtualize EMC VPlex which virtualizes IBM SVC, which virtualizes other storage arrays. You just end up in an iterative nightmare.
Worse still, what happens if IBM announces that SVC will now support the virtualization of EMC Invista Mark II (i.e. VPlex)? Now customers can watch their data circulate between two Virtualization Appliances in an infinite loop without ever again touching a storage array or server. Just caught looping in the network between appliances.
That's the idiocy of Storage Virtualization in general. It's just a way for companies to vie over control points in customer's architectures, and to get easily influenced customers to buy more (duplicative) controller and software functionality.
True Storage Federation won't need these hierachical levels of control by appliances, just as VMware doesn't need a hierachical level of appliances to enable VMotion in the server world. It just happens between peers.
Let's hope the press, blogosphere, industry and customers don't get too pulled in by all this hype. A tall order, unfortunately.