When discussing the data center of the future, it’s critical that we start by breaking the concept of the data center as a physical site with guarded walls, raised floors, neat rows of servers and crash cart pushing operators. The Data Center of 2020 (DC2020) is a distributed infrastructure comprised of many data centers, cloud services and connected devices.
The primary design concept of DC2020 is integrated automation not actual infrastructures.
As an industry, we need to actively choose implementations that unify our operational models to create portability and eliminate silos. This means investing more in sustaining operations (aka Day 2 Ops) that ensure our IT systems can be constantly patched, updated and maintained. The pace of innovation (and discovered vulnerabilities!) requires that we build with the assumption of change. DC2020 cannot be “fire and forget” building that assumes occasional updates.
There are a lot of disruptive and exciting technologies entering the market. These create tremendous opportunities for improvement and faster innovation cycles. They also create significant risk for further fragmenting our IT operations landscape in ways to increase costs, decrease security and further churn our market.
It is possible to be for both rapid innovation and sustaining operations, but it requires a plan for building robust automation.
The focus on tightly integrated development and operations work is a common theme in both DevOps and Site/System Reliability Engineering topics that we cover all the time. They are not only practical, we believe they are essential requirements for building DC2020.
Over this week, I’m going to be using the backdrop of IBM Think to outline the concepts for DC2020. I’ll both pull in topics that I’m hearing there and revisit topics that we’ve been discussing on our blogs and L8ist Sh9y podcast. Ultimately, we’ll create a comprehensive document: for now, we invite you to share your thoughts about this content in it’s more raw narrative form.