Whatever.
I've never met a CTO (or CIO) that groked technology.
Useless article. IMO, of course.
EMC Chief Technology Officer John Roese has all-but-demoted the company's Symmetrix VMAX arrays from their position atop the company's prow, today declaring the arrays are the biggest and baddest offering in a “capacity tier” of products that possess lesser grunt than a “performance tier” of products. Speaking at the EMC Forum …
They are all at it, trying to position the old flagship stuff as still relevant in the face of new entrant flash arrays where you can get good performance without a specialist machine to tune and spin it over loads of spindles. Yes there's the capacity thing, but then you still don't need the expensive flagship beast, just a bit bucket.
The solution (to storage vendor revenue cliffs) apparently is flogging software licences for virtualising access to all this.
Wow, storage is dull.
Yes, vendor panic, but not because of flash arrays, but because of applications that no longer require a SAN. EMC mentions using VMAX for HDFS, which hints at the real threat. HDFS doesn't need an array, and works better without one (but I guess if you have an old VMAX gathering dust somewhere...).
The advent of highly distributed applications like Hadoop that manage direct attachedl storage transparently, and scale processing and storage as nodes are added, is the real threat to SAN vendors. And Hadoop is just the first generation of this technology. In few years, SANs will be relegated to legacy business apps.
I recall a cloud vendor that had 2,000 servers with 6 direct attached SAS disks. NetApp told them they were crazy to manage direct attached storage, and they should with a SAN. So the cloud vendor pulled up their combined IOPS from Ganglia, and said we need a system that can handle this. Even NetApp's largest system couldn't support it, let alone the enormous costs. Arrays of servers will replace specialized SANs.
Yes, VMAX is a enterprise grade. But in this day and ages it is the SW silly. EMC's CTO is doing the right thing by positioning VMAX as a large scalable capacity tier.
Like mainframes there will be a legion of IT shops that will buy VMAX no matter what (they don't care about marketing messaging). Others who are more likely to influenced by marketing aren't going to buy VMAXs at $5/GB for HDD speeds. John's comments are paving the way for VNX2 and XtremeIO with their rich SW functionality to be positioned as the performance tiers.
How can anyone consider VNX2 and XtremeIO to have "rich software functionality?" The first is so far behind VMAX in its software functionality, implementation for implementation, that the comparison is laughable. VNX (and VNX2) is only about "acceptable functionality at a moderate price." The second hasn't got any software functionality at all, XtremeIO is only about "performance and handling Flash" (which apparently neither VMAX nor VNX can do properly). And the EMC CTO says that VMax is about "a capacity tier" (with rich software functionality, of course).
Slightly embarrasing that EMC can't provide a single platform that combines the ability to handle performance, capacity, rich software capability, and a moderate price. Not really very customer-centric. Can they keep winning market share? There may be a reversal soon. I'll be watching the Gartner and IDC numbers (and why has EMC hidden the VNX revenue numbers anyway?).