Here's another little technical preview that was quietly released during VMworld Canne 2009. To be honest I was surprised it didn’t create more of a ripple than it did. This week I filled in a survey on the Beta/RC programme for vSphere4 about the importance to me (or rather VMware customers) of this. For me it is very significant and very important. Firstly, I want vCenter to be in Linux so it can be downloaded like any other virtual appliance, decoupling it from its MS roots is the first step in that process. Secondly, it will encourage VMware customers to virtualize vCenter, which is something I’ve been recommending for years and years. I’m often shocked by the number of customers who persist in running vCenter on physical. I was shocked by the numbers of people who run on physical (60%) and there were plenty of hands up at the London VMware User Group last week that reflected this level. If you run on physical, vCenter will be limited with what VMware can do for you. You’ll be lumbered with this vCenter Heartbeat Service (aka NeverFail). My attitude is this-- if VMs are good enough for your end-users why aren’t they good enough for you? Thirdly, I would like to see VMware abandon its commitment to Oracle/SQL and single point-of-failure backend. I would love to see some kind of open-source LDAP based multiple-master model for vCenter, just like we have with Active Directory, a truly “federated” vCenter structure.
All this adds up to a vService or vApps that can be
This was first published in March 2009