Michael Warrilow at Hydrasight makes a strong case for why management tools need to support heterogeneous hypervisors. His strongest argument, in my mind, is the affinity of specific jobs to specific hypervisors. An example is Oracle’s database support only on Oracle’s version of Xen. In the realm of virtual lab management and the application lifecycle, the affinity of specific jobs to hypervisors couples with the natural heterogeneity of the production data center environment to make support for such heterogeneity a key requirement for replicating production application configurations in a virtual lab. We also believe that the ubiquity of Visual Studio as a development environment in the enterprise will result in broad deployment of Windows Server 2008 in development – which will give Hyper-V a strong initial foothold and will create new opportunities for using virtualization in unit, integration and functional testing.
What do you think?
Link: The inevitability of heterogeneous hypervisor management
I spoke yesterday to Cameron Sturdevant at eWeek regarding virtual labs and Surgient’s upcoming release schedule. One of the topics we got into was the struggle that IT operations teams seem to have reconciling their current chargeback mechanisms with a more dynamic virtualized environment. I suggested that RAM/hour or some other memory capacity-based approach worked best. Cameron wrote about it on his Permit/Deny blog.
Dan Kusnetzky at ZDnet talks about virtual lab management and Surgient as a precursor to lights-out fully automated datacenters.
Tarry Singh interviews Surgient’s Sameer Jagtap about the virtual lab management market, Surgient’s platform and product and other news from VMworld Europe 2008 in Cannes.
On Microsoft’s Startup Zone, Yi-Jian Ngo has a good post on why its important for both vendors and customers to focus on heterogeneous virtualization support in technology areas such as virtual lab management.
From the standpoint of virtual lab management, the fact that 55% of the respondents are trying to accelerate their ability to provision computing resources to end-users is a great sign. It maps to what customers are telling us – they are seeing tremendous returns from automating the provisioning of virtual machines for support, training, testing, proof of concept, evaluations etc. Those conversations suggest that a virtual lab environment is becoming a platform supporting a host of non-production activities that traditionally have been slow, expensive, and considered tactical distractions for the IT ops group. Putting an automated utility in place allows the operations team to spend their time on things that are strategic to the business, instead of being interrupt-driven with deployment and configuration requests.
At virtualization.info, Alessandro Perilli presents a very nice wrap-up of what was predicted for 2007 versus what really came to pass. One of his comments is that analysts were wrong to suggest that automation of virtual environments would play a strong role in managing virtual resources in 2007. While he’s right that it was not “key” in 07 and certainly he is correct that most virtual resources are not managed in an automated fashion, I really disagree with the overall characterization of automation as unimportant. Certainly much of the buzz currently around virtualization is centered on the automation of virtualization (be it site recovery, orchestration or virtual lab management). Beyond that I believe that automation is a buying impetus as well. I guess we’ll see in 2008!
