If you were at EMC World, you saw it everywhere you went.
VMware this, VMware that ... you'd think that maybe we had more than a passing interest.
Well -- yes!
VMware Highlights From EMC World
Chad Sakac tells me that about 20% of the several hundred sessions at EMC World had "VMware" in their title. Anecdotally, about 50% more had at least some coverage of VMware considerations in their material. Among the 9300 or so attendees, VMware was Topic #1.
BTW, Chad's a great guy -- a real powerhouse if you've ever met him. I think he should be blogging, but in the meantime, you can see him on YouTube doing all sorts of useful VMware topics.
At the show, there were great demos of things like SRM integration (including a live, no-net demo, on VDI no less), Smarts ADM working closely with virtualized stacks (including the must-see IT Compliance Analyzer component), and more. Lots and lots of customers with really intelligent questions about things like backup, and management, and -- of course storage -- in their VMware environment.
You might have thought you were at VMworld ... ;-)
Stephen Herrod of VMware's Keynote
Another reason I should have been there.
I saw the video, and got the slides -- it was a great session. Dr. Stephen Herrod -- as usual -- had a lot of very interesting perspectives.
First, he was unambiguously clear that ESX can *really* scale, as far as I/O is concerned.
He presented the results of the joint scalability testing work done with EMC, and offered up a number of 100K IOPs, which he compared to 200,000 Exchange uses, or 150 4-way databases (later corrected to 85, but still a big number).
Just to be clear, they were able to outrun what EMC brought to the party: three CX-80 arrays (no slouches) and a total of 500 spindles. Go see the whole report here, if you haven't seen it.
Next time, maybe we oughta bring a bunch of enterprise flash drives, and see what ESX can really do ;-)
Bottom line -- the "I/O scalability myth" is just that -- a myth. Maybe you have other excuses for not putting your heavy duty stuff on VMware, but don't use this one anymore.
Second, he coined an interesting twist on words: "document management for virtual machines". I found that fascinating.
Why? As we all know, with VMware, every compute image (desktop, server) is a file. And when you start thinking about many of the workflows associated with IT, they become -- well -- document management-type problems on VMDK files.
He used VMware's Lifecycle Manager to illustrate the point, which can be thought of as a workflow starting with a server image request, which is approved, deployed and intelligently placed, and -- when no longer needed, decommissioned and archived. Very much like a document.
Same general line of thinking applies to VMware's Lab Manager and the newer Staging Manager.
Although I don't think anyone is storing VMDK files in Documentum quite yet ;-)
Finally, he hit on the same theme I've been espousing for a while -- VMware has a play in the cloud.
He used his own words to express the same concept, but -- once computing is encapsulated -- it moves very easily. Hence its attractiveness for disaster recovery. And high availability. And dynamic load balancing.
And maybe -- before too long -- shipping it off to a service provider who can run it on your behalf ...
And Finally ...
If anything, EMC World confirmed the obvious for me.
Virtualization is changing the economics of computing -- permanently.
Just about every piece of supporting infrastructure (servers, storage, management, backup, security, networks, etc.) has the opportunity to be optimized for this new world. And that's what we're starting to see.
Sure, the technology is cool, but process and mindset change by IT and the business will be required to fully liberate its true potential.
And -- just maybe -- we'll start living in a world of frictionless IT.
Comments