Is It Time For Fleet Management For Applications?

Consolidating and shuttering many data center sites into a few is a high-level concept that is simple to grasp. Cost savings in the form of lower electricity consumption, fewer facility-based charges, and fewer staff are but three compelling efficiencies. In practice, meaningful data center consolidation is painfully slow as the U.S. Federal Government and many U.S. State Agencies have discovered.

One aspect affecting progress is how little is known about the application inventory across the geographies for data centers targeted for consolidation and how much effort is required to find out.

Like unregistered vehicles on a roadway, applications spring to life with no registration requirements. They operate for years or even decades and little information is consciously kept about their owners, their users, the true dependencies they have, or the resources they consume.

Some may operate like empty busses on inefficient routes until their uselessness is discovered.

It’s like our information technology is missing an entire discipline — fleet management for applications.

Some progress has been made with the advent of application dependency mapping and visualization tools. These speed the discovery process and allow a systematic approach. Not all firms can afford these tools and much of the application discovery still requires manual methods by experts. Even when the dependencies are finally understood, they still must undergo interpretation and analysis for disposition decisions.

How difficult is it to migrate the application? Which parts of the business are affected? What downtime can be tolerated? Is there a modern equivalent replacement? Who will oppose the migration or retirement decision?

Extraordinary effort is expended in discovery, analysis, and execution for ordinary (or even disappointing) results for data center consolidations.

It might be time to change tactics and start actively managing applications from the moment of their birth. It might be time to recognize the potential mobility that applications could possess instead of treating them as a once-in-a-career event to be managed as part of a data center move.

Leave a Reply

Your email address will not be published. Required fields are marked *