All Things Virtualization 
The flurry of incoming inquires on Vblock hasn’t let up yet. The top two questions I consistently get are:
- The Vblocks look like they are sized based on the number of virtual machines. Is that per Vblock? These things are modular….right? What if I stack up 2 Vblocks? 3? 10? How many virtual machines can I run?
- When Cisco, EMC, and VMware sized the Vblocks, what workload did they choose? We assume a “mixed workload,” but every company is different and it will impact consolidation ratios. How about a VDI only or a SQL consolidation workload?
- How do I justify the price premium? Since Vblocks are new and they have not run in a data center yet, how do I measure the operational impact and how can I communicate that to an executive level?
Besides that, I keep getting a bunch of people naturally reacting that Cisco, EMC, and VMware are crazy and it will never work. I always find it entertaining to see how top tier companies, sports teams, and even consumer products always seem to have a giant red target painted on their back. You work so hard to get to the top and then realize how lonely it is–and everyone is throwing stones at you.
Acadia, Cisco, EMC, and VMware: we are ready for the answers to the above questions.
Related posts:
blogs
Hi Mark:
Glad to answer these as best I can.
Q1: The Vblocks look like they are sized based on the number of virtual machines. Is that per Vblock? These things are modular….right? What if I stack up 2 Vblocks? 3? 10? How many virtual machines can I run?
A1″ You are right, they are sized by aggregate # of intended VMs per Vblock. All resources act as a single pool. Customer may need/want multiple pools.
“Stacking them up” would require a very tall data center, indeed. I think the question is “what happens when I have multiples”, yes?
Each Vblock model can scale reasonably well within its design point, but when you have multiples, you can manage them as one from UIM. Traditional data repliction will be replaced by a sexier scheme next year, stay tuned.
How many virtual machines can I run? On the big ones, a heckuva lot — see announcement for details per model. Insert usual disclaimers around “your workloads may vary” per the norm.
Q2: When Cisco, EMC, and VMware sized the Vblocks, what workload did they choose? We assume a “mixed workload”, but every company is different and it will impact consolidation ratios. How about a VDI only or a SQL consolidation workload?
You’re right, we did our best (based on customer experience) to size’em based on what we saw people doing in large-scale VMware environments. We’re working on a VDI-only characterization, as that’s a useful situation.
A SQLserver-only workload would have to be One Big Honkin’ SQLserver implementation to consume an entire Vblock, yes? We have characterized (and continue to characterize) SQLserver, Oracle, SAP, Exchange, SharePoint and a few other targeted environments as part of a “mixed workload”, because that’s the idea — a private cloud that does lots of different things.
Q3: How do I justify the price premium? Since Vblocks are new and they have not run in a data center yet, how do I measure the operational impact and how can I communicate that to an executive level.
Mark, you’ve been smoking the vendor FUD pipe I think. There is no premium for consuming the products as a Vblock vs. buying them normally. We’ll even put a reference architecture doc in people’s hands, and say build your own if you like.
All of the technology ingredients have been in the marketplace for a while, if you think about it, including “running in a data center”. Well, the UCS is a bit new, but that’s about it. Did you know that?
Communicating the operational (and strategic) impact of this approach is fairly straightforward. We put the capex and opex metrics up against what the business is getting today (not to mention the flexibility and responsiveness of the environment), and most people are inherently intrigued.
Hope this helps!
– Chuck
Social comments and analytics for this post…
This post was mentioned on Twitter by markbowker: [Blog] Acadia, Cisco, EMC, VMware: Vblock Top Three Questions: http://ping.fm/19wAH...
[...] to have the industry talking about radically new approaches to simplify computing! Here is a great post summarizing Vblock from Mark Bowker @ Enterprise Strategy Group. Now if we can only get access to that lab and get [...]
[...] This post was originally published on Liquefying IT. All views and opinions expressed in ESG blog posts are intended to be those of [...]
[...] Also read Acadia, Cisco, EMC, VMware: Vblock Top Three Questions by Mark Bowker [...]