If you're interested in a blast from the past, continue reading, otherwise check out VMware's latest solutions here!
Finally back from the PTAB meeting in Palo Alto at the mother ship. Great meeting, but I’m not a huge fan of 6 hour flights each way. There were several great discussions over the two day meeting including, of course, a very robust discussion on the new vSphere licensing model. I was able to get some clarification on outstanding questions in a few areas. First, there won’t be any true technical “tracking” of the vRAM pool(s) within Virtual Center. This means that it is somewhat honor system based, but more important, it means that if you attempt to power on a VM and you have no vRAM remaining in the licensed pool, it will still function. Think of this as a “bursting” capability, so you can go over your vRAM pool limit although obviously you’ll need to purchase an additional license to add the appropriate vRAM to remain legit at some point. Additionally, supporting infrastructure virtual machines such as vCenter itself, Nexus 1000v VSMs, vShield, and other 3rd party virtual appliances will NOT count against the vRAM pool nor will a ”standby” VM running via Fault Tolerance (FT). This is extremely important as we are seeing more and more infrastructure components like firewalls, load balancers, DLP scanners, etc. become integrated into the virtual infrastructure by means of a virtual appliance. It is very conceivable that these could eat up a significant amount of vRAM. It’s great to see that VMware realizes this and will make exclusions appropriately. As I mentioned, there was certainly a robust discussion around the impact of the new license model to our customer base. Some members are saying that they were seeing pretty big negative impacts with others saying they haven’t really seen any big issues. The customers who have done everything ”right” and have very high consolidation ratios will be the most negatively impacted, in addition to the classic Lab Manager type test/dev/QA environments where features like memory overcommit are commonly used.
On the desktop/VDI side of the licensing fence, VMware is changing its stance in the area of license migrations/upgrades. If you read the currently posted Licensing FAQ PDF document for vSphere for Desktops, it clearly states that there is no migration path for customers who have purchased traditional vSphere socket licenses to support 3rd party VDI environments (XenDesktop, vWorkspace, LeoStream, etc.). However, this will not be the case, and VMware will in fact be offering some type of migration for these customers into the new vSphere for Desktops model. I don’t yet know any details as to how this will work, but good to know it is in process.
I was also able to verify that the new SRM licensing model is what I thought it was. There is zero functional difference between the Standard and Enterprise edition of the product. If you have 75 VMs or less within a single datacenter that you wish to protect, then you fall into the Standard category; anything more than that will put you into the Enterprise edition. It doesn’t make a ton of sense to me to essentially penalize larger environments ($195/VM for Standard, $495/VM for Enterprise) rather than offer the larger shops a volume discount but guess VMware doesn’t see it that way.
That said, this is really positive news for the smaller customers as most DR solutions tend to be out of budgetary range, and we often end up architecting a substandard solution simply to meet a budget number and provide ”something” vs. the right thing. Now, smaller environments can benefit from a very robust automated DR solution at a very reasonable cost, so I look forward to having more conversations with GreenPages’ customers in this area in the near future.
Obviously, there was lots of conversation around the infamous cloud, and Paul Maritz himself spent about an hour speaking to the group outlining his vision of the future and answering our questions. I thought it was fantastic that Paul took the time to speak to us and get some feedback. He absolutely ”gets it” and I believe has a very solid vision for where he wants to take VMware as an organization over the next 3-5 years. His vision includes much tighter integration between the core VMware stack and making sure that vCloud Director ends up being that magical single pane of glass that everyone wants. Additionally, other recently acquired technologies like Socialcast, SlideRocket, etc. will be integrated where it makes sense. I think one big challenge VMware faces in the short term is communicating Paul’s vision and message out to customers as well as to the masses within VMware; many of the people I’ve spoken with inside the company were not able to outline the vision as well as Paul did. Wish I had a recording of that to share, but I’m sure the messaging will make it out over time. In the end, my interpretation of his vision is that he wants to make the operating system completely irrelevant on the desktop, as well as the server. Project Horizon is the foundation of this, so expect to see a lot more hype around it during VMworld. Horizon is essentially VMware’s version of an app store where the applications will be delivered by whatever means make the most sense based on where you’re physically located and the type of device you have at the time. For example, if I’m on a laptop and need to launch Outlook or Word, it would use a locally installed copy whereas if I did the same thing from an iPad it would launch a VDI or terminal server/XenApp type session. In the end, I get the application I need without really having to think about or care where it comes from--the essence of cloud computing.
Unfortunately, I won’t be able to attend VMworld myself this year as I’ll be on vacation in Europe ( I know, poor me), but I look forward to hearing all of the news when I return. In the meantime, I’m very interested to hear the community’s thoughts on this news.