Virtualization and green computing
"Green computing" seems to be seems to be the up and coming thing for suppliers offering virtualization software as well as management software for virtualized environments. It's a topic that has come up over and over again in my discussions with suppliers recently.
What is "green computing" you ask? It appears that there are many related, but different, definitions floating around the Internet. Most of them focus on a set of processes and approaches designed to make datacenters more efficient. This means reducing the power and cooling required. I have read that the daily power consumption of a typical datacenter is equivalent to the monthly power consumption of thousands of homes. How many homes mentioned typically depends upon the size of the data center, the number of systems in the data center and how large of a check the supplier is hoping your organization is going to write for their product. This has become a large enough topic that the U.S. Environmental Protection Agency has started a study of datacenter power consumption.
What does this have to do with virtualization?
As it turns out, it has a great deal to do with the subject. One of the primary goals of almost all forms of virtualiztion is making the most efficient use of available system resources. In a series of future posts, I'll look at each layer of virtualization technology and consider how it contributes to green computing.
For example, many of the suppliers of management software for virtualized environments have the ability to move workloads to a smaller number of servers when workload and service level agreements allow it. Then the unused servers can be powered down. Cassatt, DataSynapse, Novell, Scalent and Surgient all mention this when discussing their products. I'd like to note that making efficient use of IT resources though automatic management of resources has been part of Novell's marketing messages for well over a decade if my memory serves me correctly.
What is your organization doing to make their operations more efficient? Would your organization deploy tools such as Novell's Orchestrator, Cassatt's Collage or DataSynapse's MatrixServer in the hopes of reducing power consumption and heat production while still being able to meet established service level agreements?