Last year I wrote a post on VMTurbo and its method of using the idea of a market economy to manage your infrastructure. If you need a refresher (or because you didn’t read my blog, shame) take a look here. If you aren’t going to read it, the gist is that VMTurbo monitors your virtual environment and uses the hardware as though it is a supply, and the workloads that run on it as the demand. Based on the demand of a workload and supply of a resource there is a cost associated with the workload, and VMTurbo uses these metrics to determine the most cost effective way to balance these.
All travel expenses and incidentals were paid for by Gestalt IT to attend Virtualization Field Day 5. This was the only compensation given and did not influence the content of this article.
Fast forward 18 months and I’m back at Virtualization Field Day 5 where VMTurbo is expanding on this idea. Hybrid Clouds are picking up steam from customers and users are actually using both at the same time in their datacenters. This affects the “Global Economy” that VMTurbo has been putting together. To simplify this concept, assume that we have a regional market like the United States (Private Cloud example) and we’re moving workloads around within that country’s borders to make sure our workloads are cost effective. Now we’ve just got a trade agreement with France (Public Cloud Provider example) and they’ve got some unlimited cheap resources! VMTurbo would now manage both of these clouds and determine where the resources should run in the most cost effective manner.
Well, I just said that France (Our Public Cloud Provider) has unlimited cheap resources, so VMTurbo would obviously want us to run all of our workloads there right? Not necessarily! VMTurbo has baked into the platform an additional cost for transit of the goods and services. Think about it this way. If France has lots of cheap resources but we have to ship them back to the US to use them and that cost is more expensive than getting resources from the US, would we do that? (The answer is no here.) To handle some of this, VMTurbo can monitor the networking stack using Netflow and analyze where workloads should be placed. VMTurbo is calling groups of related applications a vPOD. VMTurbo’s software will make a recommendation to keep these resources close together in order to better utilize the transit networks.
VMTurbo is also introducing QOS for applications as well. The idea behind this is the applications that have a SLA set on them will be treated as a customer that provides a higher revenue to the company. That application would get some preferential treatment since it’s so important to the economy.
I really love VMTurbos basic premise of using resources as a supply and workloads as a demand. This accurately sums up what is actually happening. Much like our economy changes quickly and drastically, so does our infrastructure. VMTurbo is making modifications to handle these changes and I hope they continue to build on this economy idea.