Who knows and to be honest who cares.
They're both the same incorrect approach for Enterprise workloads, they both brain damage the arrays you put them in front of as they act as external caches, and they're both seriously performance limited.
What's telling about this latest entry, besides how many different kinds of crazy is now coming out of IBM's storage division these days (The less said about IBM's breathless announcement that they now sell more DAS & Tape than EMC the better), is Tony's assertion that EMC is somehow behind spreading that information around. For those who like mysteries ignore this interview with the DataCore guys on the same site, written by the same reporter, posted one day after the rip off blog entry.
While it's an interesting delusion I'm afraid that at the frequent League of Evil meetings, where EMC works hand in hand with other dark powers of the universe to make the world a rotten place, the DataCore connection never came up. Or maybe I missed that meeting if it occurred during the annual baby eating contest EMC runs internally. Even evil has to unwind sometimes.
Tony then goes on to try and paint Invista as some product customers put together & tune themselves on site (They don't, nor do they buy their own x86 servers for the Control Path Clusters like he implies) as opposed to the IBM way of adding SVC nodes (Which are also x86 servers, though maybe IBM's are somehow magic?) until you just run out of nodes you can add.
I think what's needling Tony, besides having a product line so stale it exhales dust, (SVC and resold NetApp gear being the only interesting products they have across their entire portfolio), is Invista's sudden appearance over his shoulder.
SVC is the only sexy storage product IBM have had to talk about for a long while so they were hyper focused on it while EMC was working in the background trying to do things in the Fabric while cleaning IBM's clock everywhere else in the networked storage business. But then an interesting thing started happening. Irate SVC customers started popping up just before the Invista 2.0 announcement complaining about the performance of their production SVCs not meeting the expectations they were assured would be met, and all of a sudden doors which weren't open during the 1.0 release were opening for Invista 2.0
IBM take an early lead in a market only to screw it up and then watch as EMC takes it from them.
We've all seen this movie before. We'll all see this movie again.
The Invista 1.0 release was a 1.0 product and was treated as such by customers. Lets face it 1.ohhs are always a hard sell unless you're throwing it in for free with something else in order to seed the market. IBM like seeding the market in their storage business in the same way they like to shout about the number of SVC nodes shipped like its the number of SVC customer deployments.
Those nodes are deployed all right, they're just deployed in the same frame at the same site and are still incapable of scaling anywhere close to a single Invista instance.
Invista 2.0 with RecoverPoint is a much more interesting value proposition for people as it starts doing a whole lot more with those intelligent switches. You start putting more and more heterogeneous services into the Fabric and getting access to more and more functionality which isn't going to lobotomize your existing infrastructure.
It took a full product revision but it finally feels like Invista has traction, I expect interesting things from here on in.
And I expect even more hysterical flights of fantasy from Tony.