By Sramana Mitra and guest author Shaloo Shalini
SM: Okay, that gives us some context. Now let’s explore what your philosophy for computing is and where HMH is in terms of cloud computing adoption. Have you moved on from the pilot stage to broad deployment, and which workloads have you moved to the cloud?
PW: Almost four or five years ago, we made the decision as an organization to consolidate all of our external hosting providers into an owned data center in Wellington, Massachusetts, just outside of Boston. We did that and combined our internal applications, our internal support applications, and our externally facing technology products. When we did that, we first conceptualized that data center as being a private cloud. So, we took a shared services or infrastructure as a service approach to creating a data center that has a capacity-on-demand model and got away from the one application, one stack of hardware mindset that we had at the various hosting providers. I think that set us up well from a private cloud perspective, and we continue to operate down that path and resist polluting the data center with a dedicated hardware sort of an approach. Our approach is more of the shared infrastructure offering layers and support to products and other applications.
SM: Great. I have a number of follow-up questions on what you said about your private cloud strategy. A lot of CIOs say that they don’t want to be in the data center business, and one of the reasons they are adopting cloud computing so aggressively is that they want to get out of data service. But it sounds as though you have chosen to make being in the data center business part of your strategy. Would you talk about that?
PW: Yes, it is sort of funny, and I guess there a couple of things behind it. I am by trade and education a finance person, not a CIO. For 18 years I was in banking and finance and the management of financial institutions. So, I have a pretty deep financial management background. When I came to publishing 15 years ago, I came into the K-12 educational division, so I have been in the business and I have seen the way we went about creating technology products. It was always a one-off: you went out and found your solution with a hosting provider, you created that environment, and the next time you needed another you did the same thing. The redundancy was tremendously costly.
There was no virtualization whatsoever. It was very much a dedicated alignment, even though I kind of smiled when I looked at the first page [of your] questionnaire where there was a question about the hype cycle. Being around technology as long as I have, you could put a lot different technologies into that cycle, and to my mind we have always had some sort of cloud computing service; you just didn’t call it that. You just said, I need storage, or I need an application, or I need an application server or a Web service, or whatever. Now, the products and offerings are bundled together into the concept of the cloud, but when I really did the financial analysis and looked at what it was costing us to do these sorts of things, separately and in an on-demand way, the cost was huge. When we actually built the data center and consolidated into it, the investment paid itself back in less than a year.
So, financially it was feasible as we developed the layers of servers: Web services, application services, database services, and gave folks a bit of a choice about what they purchased, if you will. It gave us much stronger governance, it gave us the ability to consolidate and virtualize, and it also gave us a much higher level of reliability because we were not managing multiple, disparate services.