Large Pages, Transparent Page Sharing and how they influence the consolidation ratio
This is a great article by VMware expert Gabri van Zanten. It also is a great example of why I think Dynamic Memory is going to be a very popular feature.
In short, Gabri rightly states that using Large Pages is a tradeoff between memory savings and performance benefits. He also doubts if the use of Large Pages really bring any benefit to your application. I think that in VDI environments people want to maximize the use of their RAM without it having to be an exact science that people like Gabri have to write multiple articles about. As far as I have seen, Dynamic Memory does the job fine in a fairly transparent, ‘real-time’ way. It also reduces the risk of excessive swapping and prevents the risk of kernel mode RAM to be swapped out.
Anyway, still a great article to read by Gabri:
Everybody talks about the benefits of Transparent Page Sharing in vSphere environments and how they reduce the amount of memory needed in your vSphere host. Thanks to Transparent Page Sharing memory overcommit in a production environment has become mainstream.
Not long ago I’ve written these two posts:
- “Memory overcommit in production” http://www.gabesvirtualworld.com/memory-overcommit-in-production-yes-yes-yes/ which explains how Transparent Page Sharing works and how to use it in production.
- Another post of mine explains more on memory compression and how ESX starts ballooning and swapping when there is memory contention, see: http://www.gabesvirtualworld.com/memory-management-and-compression-in-vsphere-4-1/.While writing that post I started thinking about the impact of Large Pages and Transparent Page Sharing on memory usage and how this would make it a more difficult in your day to day admin job to monitor your real free memory on the ESX host. This could therefore lead to less VMs per host.
Let me explain what the problem is.
January 26th, 2011 - 17:01
I don’t think Dynamic Memory (hot add/remove memory) can be compared to TPS which collapses pages. On top of that and especially with VDI think about what happens when a host is under memory pressure with Dynamic Memory. VMware doesn’t only offer something similar (ballooning), but on top of that TPS and memory compression.
January 27th, 2011 - 10:54
Hi Duncan,
I am not trying to say that TPS is the same as Dynamic Memory. The point I am trying to make, is that the end goal is/should be the same. For VDI the most important thing in my mind is to maximize the use of their RAM without it having to be an exact science. I have seen that DM delivers this.
I also know that VMware’s memory management techniques deliver this as well but at a more complex configuration (to me:Windows guy) and with a higher probability of overcommiting memory and the risk of swapping kernelmode memory.
I would love to know how the two memory management techniques compare in a VDI benchmark but I think the difference shall not be great. What do you think?