Elasticity in the Mainframe Environment

By Brenda J. Christie

This week's Enterprise Systems Media feature article centered around the concept of "elasticity," a nice buzzword that really means flexibility.  Or in project management speak, it is called "resource balancing."  However, the article, "IT Sense: Elastic Storage" talks about the difficulty of assessing the availability of a resource due to a lack of standardization within the environment itself.


The article begins by discussing the desirability of having all human resources ideally active in order to minimize paying for a resource that is not being used.  The idea is to be able to use software to monitor output and then re-allocate idle or under utilized resources to a task which could use more resources.

The article then moves from discussing human resource utilization into discussing detecting available capacity within the storage arena, and attributes failure to do so to the fact that not all storage is created equally, and configurations vary due to cost and capability.  In essence, it says the concept of Elasticity is not viable.

Well, I have to disagree with that tenet.  To compensate for the differences in storage, meters could be attached to the storage devices to capture output levels.  Given the existence of pre-defined benchmarks as to optimum throughput, it would be possible to tell whether a device is at, under, or over capacity.  Assuming all the meters are somehow connected to one control device, it should be relatively easy to tell which storage device needs more work and what the new optimal configurations should be.  Not rocket science.

Indeed, industries outside IT have already conquered this problem.  Witness MotorsAtWork.  As part of my work in property management, I have seen meters used in commercial energy consumption monitoring and performance tuning and will post a follow-up link to an article highlighting such successes.

Brenda J. Christie