I’ve just wrapped another long day at VMworld Europe in Cannes, France. I’ve got lots of thoughts on many things I’ve seen and heard here, but I thought I’d start with a topic that Brian Madden brought up in his blog about how VMware is misleading everyone on TCO.
Brian saw the partner-facing version of the VMware View TCO session on Monday – I saw the same basic session repeated on Tuesday for the general customer audience. By this time, the VMware representative must have read the blog and specifically noted, “This is not intended to compare View to other alternatives, like Terminal Services.” Okay, cool.
Now the part that I have contention with is a part that I actually agree with VMW 100% on – let me explain. In the session, she talked about how the cost and ROI analysis should be broken into 3 buckets – capex (infrastructure acquisition costs), opex (support, management, adds/moves/changes) and then end-user costs/benefits. Rightly so, she point out that end-user costs/benefits are hard to quantify so approach with caution, and for the most part, the case for VDI will not be made on capex savings. That leaves the bulk of the savings being generated from efficiencies in Tier 1/2/3 support, application management, etc. Oh, and by the way, she repeatedly mentioned, Gartner put out a report saying that the TCO savings in this area were up to 50%. WHOAH!!!!
The slide was comparing traditional managed PCs to VMware View, and Gartner said you could save 50%?? That would be big news, wouldn’t it? Wait a darn second… I get every Gartner report on the space and I don’t remember seeing anything like that….
Turns out that the Gartner report being referenced is entitled, “Total Cost of Ownership Comparison of PCs With Server-Based Computing.” If you are at all familiar with Gartner taxonomy, you know that VDI equates to Hosted Virtual Desktops (HVD), and Server-based Computing (SBC) equals…… Terminal Services (TS) and XenApp. If I’m not mistaken, there’s no TS or XenApp in VMware View (not seriously, anyway), so that leaves me to assume that VMware think the TCO case for VDI can be made by taking the numbers from a completely different solution architecture.
In Brian’s “VDI vs. TS” session at VMworld (bravo, by the way, great session), he made a good case for VDI and TS to both be considered SBC. I buy that there are some inherent benefits that apply to any centralized, server-hosted infrastructure, so some generalization might be appropriate. But Gartner get paid the big bucks to be VERY precise about taxonomy and TCO. They factor in things like server density, app patch management, etc. for SBC that are completely unrelated to VDI. Munging the two together is just a raw misapplication of the research and misleading to say the least!
That said, I have to mention that I think the only misleading part is associating the savings with View – the savings analysis fits perfectly for XenDesktop! With XenApp baked in as part of XenDesktop, you indeed get the benefits of SBC as part of the solution, as well as single image management from our provisioning services, user productivity from HDX technology, etc. So I think the approach for using the SBC TCO analysis in the context of VDI was spot on, it was just done for the wrong product!