Hyper-V R2 Dynamic Memory: over-subscription vs. over-commitment

This content is 14 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

There’s been a lot of talk about how Microsoft’s Dynamic Memory capability in Hyper-V R2 compares with similar features from VMware – including the pros/cons of each approach. Because that’s been so well-covered elsewhere, I’ll avoid it here (from the VMware perspective, check out Eric Gray (vCritical),  for Microsoft Ben Armstrong is your man and, for a completely unbiased and objective view… well, good luck finding one). I did see an interesting quote however from one of Ben’s TechEd sessions in New Zealand recently:

  • “Over-subscription is what airlines do by selling more seats than places in a plane.
  • Over-commitment is what happens when all those passengers actually show up to use their seat.”

[Ben Armstrong at TechEd New Zealand]

One of my fellow Virtual Machine MVPs, Ronald Beekelaar, extended this analogy and it seemed good to share it more widely…

There is nothing wrong with over-subscription – it happens in many real-world scenarios such as: public transport; libraries; Doctors’ surgeries; hospitals; utility companies; telephone systems; etc.  – and these work well (most of the time). The issues occur when all of the people that could actually use the service try to at the same time, at which time we have over-commited the service.

What do we do when we have over-commitment? We add more resources (run extra buses, add carriages to a train, add books to the library, open a new hospital ward, lay more telephone cables, etc.) – and in the world of virtualisation, we add one or more hosts and migrate some of the conflicting workloads away.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.