Profiling Workloads

In any consolidation, abstraction, and virtualization project, it should be standard practice to use profiling tools to measure workloads over a significant period of time. Important metrics to measure include utilization of:

  1. RAM
  2. Disk I/O (input/output)
  3. network
  4. CPU

It makes sense to use a tool like Platespin Recon to measure these things.
http://www.novell.com/products/recon/

Citrix article on profiling, mentioning PlateSpin Recon: The importance of knowing your environment
http://blogs.citrix.com/2010/04/02/the-importance-of-knowing-your-environment/

Novell has a Project Edition which can be used to measure a server's workload at $2/day per physical server.
http://www.platespin.com/powerstore/

If you have a VMware partner, they'll be able to use the VMware Capacity Planner for you.
http://www.vmware.com/products/capacity-planner/

If you have an all-Microsoft environment, you might consider their Microsoft Assessment and Planning (MAP) Toolkit (free!).
http://technet.microsoft.com/en-us/solutionaccelerators/dd537566.aspx

Other options are: