How to Design a Workload-Centric IT Infrastructure
The current siloed approach to IT infrastructure is no longer sufficient or efficient. In order to deliver better business results in the coming years firms must adopt a 'workload-centric' approach when designing their IT infrastructures. Getting there will require businesses to change the way they think about people, process and technology.
Mon, October 01, 2012
CIO — Today's approach to IT infrastructure is inefficient. Your business cares about outcomes and services, but it's built on technology-centric silos. For most firms, each silo is represented by an autonomous department that selects technologies in a vacuum -- instead of realizing the value of integration with the other pieces of the firm.
In Forrester's IT Infrastructure Playbook, my colleagues and I maintain that in order to deliver better business results over the next decade, firms must adopt a "workload-centric" approach and design IT infrastructure on what matters most -- the workloads -- and not the other way around. But what exactly do we mean by workloads? We mean the major end-to-end application or data services that are the center of control and technology selection in your business environment. For example, server virtualization is a key workload because it aggregates many business applications into a common infrastructure.
Workload-centric infrastructure will remedy the ills of today's siloed approach -- but getting there requires a different way of thinking about people, process and technology.
People: Align Teams With the Workloads and Applications That Matter Most
Typically, individuals are siloed around elements that they are experts on. With this inside-out approach, it's easy to miss the forest through the trees -- technologists know a great deal about the category of gear that they buy and manage, but little about how it is used and what the customers really want. Because each silo functions independently, it's hard to move in a common direction.
In a workload-centric organization, staff members need alignment with the workloads they serve. To minimize disruptions but improve focus on key workloads that matter, Forrester suggests creating a multidisciplinary task force for each key workload. If each server, storage, and network team designates one or two members to attend planning and architecture sessions around workloads, the process of improving communication across departments can begin.
For a bolder approach, firms should eliminate the technology organization entirely. Instead of taking people and funds from each department, forward-looking firms are starting to consider aligning their infrastructure with workloads. For example, consider a VMware, an Oracle, and a mainframe team each reporting to a single leader. While this is clearly a departure from current practice, the improved communication, decision-making, and ability to track the true total cost of a workload is attractive.
Process: Measure the Effectiveness of the Entire Workload
Processes and metrics are rarely shared across server, storage, and network teams. For customers, this results in a confusing list of engineering specifications and finger-pointing at other departments. Customers simply don't care where the bottleneck or error lies, they just want to know when they will be back up and running or why they aren't going fast enough.