Several years ago, I decided to earn a second Master’s degree in strategic planning from the University of Southern California. One of the things we studied was futures research. In futures research, sometimes you try to determine where things are going by scouring the points that fall off the curve.
For me, I wanted to know what IT would like five years into the future. So I asked around and found a leading-edge organization to talk to. And I went to their architects. You know these folks; they are responsible for the current state and future state of IT. I started my questioning by asking, “What do you believe you will be buying in the future?” Instead of responding “servers, storage and networking equipment,” they said that they wanted to buy “IT.”
I asked them to tell me more. A few years earlier, this organization quickly jumped onto virtualization and as a part of this, they moved some of their production infrastructure and services to Amazon. They said this move resulted in an amazing outcome, it made on the political stuff on virtualization go away. In fact, it has gotten them to being 90% virtualization—they believe now that they can get to 100% virtualization and even more striking that they do not need in the future need to be in the datacenter business. In five years, this Fortune 500 organization says they will have no datacenters at all. So what does this say about enterprise architecture? They said that it remains really important, but eventually it will be important to far fewer people; you need enterprise architecture to prevent sprawl at someone else’s datacenter.
With this, I asked about services. They said the back office will be gone and much of the front office too. They said that enterprise resource planning (ERP) does not provide competitive advantage and will be consumed via software as a service (SaaS). In fact, they believe that all packaged apps will be consumed via SaaS. At the same time, they said that the only thing they will build will be things that relate to their business capabilities system and these will be provided via infrastructure as a service (IaaS).
So I asked, “Is there anything that will remain in a company’s data center?” The architects could only think of one application where owning a datacenter would continue to make sense—trading, and in particular high-frequency trading applications, where microseconds of latency really matter.
So where do you do you think the journey toward public, private and hybrid cloud will end. Will it end in the death of traditional ERP, shared services and finally the owned datacenter itself? Please share your opinion here.