The modern data center environment has become the heart of almost any organization. Because of this, there has become a greater emphasis on creating efficiency around data center systems. This means eliminating complex distributed resource platforms in favor of optimized converged infrastructures.
One of the expensive challenges of hosting any workload at significant scale is server sprawl. For example, many organizations start off with one Exchange Server. This server eventually reaches capacity, so they add another. Then perhaps another, and then an extra because they want to ensure that they have enough capacity in the event that one of the existing servers fails.
You don’t have to look far to find signs that forward-thinking IT leaders are seeking ways to make infrastructure more adaptable, less rigid, less constrained by physical factors--in short, more like software. You see it in the rise of DevOps and the search for ways to automate application deployment and updates, as well as ways to accelerate development of the new breed of applications and services. You see it in the growing interest in disaggregation--the decoupling of the key components of compute into fluid pools of resources.
DevOps is a new trend in IT today, and with good reason. Composable infrastructure was designed using many of the same principles as the DevOps movement, and as such is well aligned to be the infrastructure of choice for DevOps activities.
In a perfect world, IT pros correctly estimate the workload of a particular server over the course of its projected lifetime. When they purchase the hardware for that workload, the workload grows into the hardware until it fits like a glove. Unfortunately, when most workload estimates bump up against reality, they are found wanting.