To Microsoft's competitors, the impending release of Hyper-V has been like waiting for the slow arrival of a comet from space. Today, the comet has arrived, and it will undoubtedly have a deep impact. This doesn't mean that Microsoft's competitors need to go and hide in a cave. Instead, they have known about the comet's arrival for some time and have had plenty of time to prepare. So I expect Microsoft's virtualization competitors to begin to show their competitive hands in the coming weeks.
I'm not going to spend a lot of time on Hyper-V's features since I recently wrote a review of Hyper-V for Virtualization Review Magazine. Hyper-V's core technology is good. It's a stable hypervisor and it's been hosting Microsoft's MSDN and TechNet sites for months. Feature-for-feature, Hyper-V still has a ways to go to be on par with VMware. Live migration and memory over-commit are two features at the top of my list. Sure, in some instances we can get by without live migration, but I'm even sure Microsoft would admit that it cannot achieve it's own dynamic data center vision without live migration as a core component. Quick migration is close, but migrating a VM to another server will result in the migrated VM being unavailable anywhere from several seconds to several minutes. Live migration, on the other hand, doesn't induce any downtime. There's a lot of debate out there regarding memory over-commit, but from my experience I've seen it make a substantial difference with regard to consolidation densities, with typical improvements in the 40% range. Those that don't have it will say that at peak load it doesn't make much of a difference because VMs will need full access to physical memory. This is true if all VMs on a physical host realize a sustained peak at the same time, but that's rarely the case. Instead, peak workloads are often sporadic, so the VMs that need the physical memory will use it and those that don't will have a portion of their physical memory swapped to the hard disk. From a consolidation density perspective, memory over-commit will continue to be important.
Hyper-V is missing a couple of key features, but that's to be expected in a 1.0 product. Microsoft needed to get Hyper-V out, and they've done it. Clustering VMs is extremely easy with Hyper-V and System Center Virtual Machine Manager 2008 (currently in Beta), and since the hypervisor comes with the OS, there's no added cost. What does all this mean for VMware? VMware is now up against the "good enough" factor. Sure, organizations aren't lining up to run all of their production workloads on Hyper-V, but they are considering it for some use cases: development and test, training, and virtualizing the branch office. These cases often come up due to cost. Hyper-V is free with the OS, and a Windows Server 2008 license includes "downgrade rights." So for a typical Hyper-V deployment I may buy or upgrade to two Windows Server 2008 Datacenter Edition licenses, which will allow me to run Hyper-V on two servers in a cluster and run an unlimited number of Windows 2003 VMs on the cluster. What's the prevailing reason for considering a 1.0 hypervisor. Most of the client's I've talked to have indicated cost as the primary motivating factor. Also, many have liked what they've seen with the System Center suite of management tools.
I'd be surprised if VMware is going to let Microsoft win the cost war, and as painful as it may be, I would be shocked if VMware didn't drop the price of it's ESX hypervisor in an attempt to take cost out of the equation.
While the first wave of the Microsoft virtualization comet is impacting the hypervisor community, the second wave may have an even larger impact. System Center Virtual Machine Manager 2008 is currently in beta, and I'd bet that we will see it released to manufacturing in the near future. After all, much of the really good management functionality for Hyper-V is in System Center Virtual Machine Manager. System Center is where it really gets interesting. This is because Microsoft's management suite is now reaching a point where it could take market share from the predominant enterprise management software vendors. To me, this has to be VMware's play. I don't see VMware as trying to directly compete with HP, IBM, CA, or BMC, for example. Instead, they have to work to build stronger partnerships. Such an approach would give them leverage in what looks to be a long and drawn out fight with Microsoft for virtualization supremacy.
What are your thoughts on Hyper-V's impact? I would love to hear them.
Posted by: Chris Wolf
Comments