Server Consolidation, Virtualization and Service Criticality

Complexity & High Risk Is An Argument For More Knowledge Not Less! This morning I was reading my copy of Network World and pondering the trends in IT Service Management in context with the hot topics being talked about within the recent technology centric magazine articles. For several years we have been seeing a focus on technology consolidation and lowering transactional computing costs with an emphasis of doing more with less by leveraging advancements in processor, memory, storage and OS virtualization technologies. In short, there has been a massive push to reduce the complexity of IT management by reducing the number of physical locations, applications and components that enable IT services. Notice that I did not say that the number of IT services are being reduced! In fact, the trend for IT Service Management is going in the opposite direction. The business need for, and dependency on digital automation, continues to rise while the technologies that support these services are being condensed and consolidated to the point that any specific component is potentially supporting multiple IT services of varying degrees of business criticality. This fact makes understanding Configuration Management (CMDB) relationships even more critical than it has been in the past. For Example: Let's say you have a physical server (a component of your Infrastructure Service called Hosting) which hosts three different virtual servers each supplying the processing support for a business application service. However, these three application services have very different levels of business criticality. Now let's say a change has to be made to the parent physical server to address a patch level. What risk level do we assign to this change? The answer of course is that the physical server must inherit the highest risk level of the services that depend on it. This means that we maintain a clear understanding of what each component does in relationship to business outcome. The challenge to this statement is that not only is the maturity and discipline of Configuration Management very low across the entire industry but that the complexity of tracking these relationships is growing with the move towards dynamic resource allocation. In short, an application service may be moved from server to server based on automated load balancing technologies making this critical business intelligence even more challenging to obtain. Based on this observation there are two possible responses:
  1. We throw up our hands in frustration saying that Configuration Management will always be an impossible pipe dream not to be realized in our lifetime

  2. We realize that the very nature of the growing complexity and business need for this knowledge is the business case for why we need to do Configuration Management with a realization that it will require the use of IT discovery and monitoring tools as well as some good old fashion discipline and effort
Troy's Thoughts What Are Yours? ”The difference between try and triumph is a little umph.” ~Author Unknown

Like this article? Like

Comments

Post a comment