Thursday, December 18, 2008

Last week, on his blog Thinking Out Cloud, Geva Perry wrote about how he thinks that we are going to see a wave of innovation surrounding data center infrastructure in his post Infrastructure is Sexy. I have to say that not only do I agree, but we are seeing it in action right now. We have the opportunity to work closely with some of the leading infrastructure hardware, software and service vendors in the industry, and I can tell you without a doubt: the investment wave has already begun and we will begin to see the rewards of that investment in the first half of next year. In fact, I was talking with one of those vendors just yesterday, and I'm excited about what is about to hit the market. (I can't share details though, as I need to respect their NDA.)

In my role, one of my focus areas is infrastructure management software and automation. Given that focus over the years, I've been very frustrated with the level of support provided by the software vendors for service providers (and this includes ITSM applications, in addition to the automation frameworks). While this might have been an acceptable way to focus on getting into the enterprise market, trends in IT are going to drive CIO's to continue to push their infrastructure into service provider's hands. So... it's our turn for some focus from the vendors.

The shift to service providers is due to a combination of cost cutting pressure and the fact that the increase demand for infrastructure has created what seems to be a never ending stream of capital investments required to meet that demand. Shifting spend from capital to variable, with cost drivers linked to real demand, is exactly why cloud computing is exciting. Service providers have an opportunity to provide clients with different expense models by leveraging scale and the laws of averages to drive down the overall capacity requirements across a wide variety of customers.

For us to be effective at this, we have been relying on massive customizations or manual work-arounds when using the current generation of automation software packages. For the next generation of systems, my hope (and I think I may be in for a treat) is that we see more focus on multi-tenant solutions and increased programmatic access into the packages. No vendor will own our entire management infrastructure, because no vendor can really manage all of the different elements that we need to manage. It has remained up to us to sew these things together, using whatever hooks we can grab a hold of and some standards (SNMPv2 as the most basic example). The more vendors understand this reality, the more they will be willing to support our need to put these packages together in unique and scalable ways.

Labels: Infrastructure 2.0