Virtualization in Production: Flexibility the Driver, Not Cost

Yesterday, Hyperic published a case study about Mosso, a division of Rackspace. Co-founder Jonathan Bryce tells a great story. They manage 30,000 websites and applications using a sophisticated architecture that is 100% clustered and 100% virtualized using VMware.

Of course, the case study talks all about how Hyperic HQ is the software that made this setup manageable. It’s the whole point of the case study and the press release.

However, what I think is just as interesting is Jonathan’s comments on why they would develop a production system with virtualization. It doesn’t save him money, which is the usual cost driver. His servers are actually losing capacity because they have the overhead of running virtualization software. (This is not a slam on VMware, its a fact that his servers are extremely utilized as is so any software, no matter how lightweight will constrict capacity).

The reason Mosso uses virtualization in production is FLEXIBILITY. Being able to deploy extra or move capacity in minutes is a huge plus for businesses like Mosso. This isn’t the first time I’ve heard this. As Hyperic is designed to manage web applications and infrastructure, I get the opportunity to speak with operations teams from many of the largest websites in the world. Many are considering virtualization in production. Not for cost savings. That is a driver for optimizing test and dev environments. Production deployments, as it turns out are driven by flexibility oriented to drive better customer service and to quickly deploy new functionality or capacity.

This type of flexibility is important to improving customer service and experience. The process is expensive, with both the cost of the extra computing power and software. Even more costly is the overhead of keeping the levels of service with this type of non-traditional, layered environment. This is the hurdle that prevents most operations from being able to do this.

This is where Jonathan’s story gets even better. Spoiler alert: It involves HQ, with a lot of other processes he put in place to make sure he has 100% control over his environment.

In addition to the case study, Dan Kuznetsky captured Jonathan’s story nicely today on ZDNet. He also makes a call out to the Hyperic community to see what your experiences with virtualization in production are. If you have a story you’d like to share, you can comment here or on his blog here.

If you like this post then please consider subscribing to our RSS feed. You can also subscribe by email and have new articles sent directly to your inbox.