The wheat and the chaff – do you really need all those capacity metrics?

For a Capacity Manager or Analyst, there is a mass of data typically collected, but very little that ever needs to go forward to others.  At a component level, many organizations now have Capacity Management for thousands of servers, petabytes of storage and more, vested in the hands of just one Capacity Analyst.  Pulling all this together and selecting the ‘zingers’ to pass forward to management is a monumental challenge. 

One capacity study we were involved in as consultants covered the merger of two organizations each with revenue in excess of $5 billion.  Multiple data centers, thousands of servers and a request from management to evaluate their plans to merge systems and reduce the hardware estate.  Management wanted to know when in the migration plan there were pressure points where resources might be stretched and additional investment required.  Their preference was for this to be communicated within a broader presentation covering other issues, and be communicated in just one PowerPoint slide.  It’s not an unreasonable request: tell me when a decision needs to be made, and what that decision should be.

The quality and skill of those involved in Capacity Management never ceases to amaze me.  Answering demands like the one above is regularly achieved.  For us, the key was understanding the end point.  Once we knew what they wanted on that one PowerPoint slide, we could focus our efforts much more precisely on what data we needed to get and how we needed to use it.

We would of course recommend that the wheat be separated from the chaff through third party products like Metron’s athene®.   For many businesses, that is a step they have not yet taken.  Questions are addressed by patient manipulation of raw data and use of tools like Excel.  One organization we spoke to recently had Excel macros producing over 5,000 individual Excel spreadsheets every week, just for their UNIX systems.  Then there was Linux, VMware, Windows…..  

Whatever works is fine, although each approach has its risks.  In particular extensive home grown routines are susceptible to those who have created and know how to maintain them moving on to pastures new.  Third party solutions at least have the benefit of external support enabling easier transition of reporting and analysis regimes to new staff.

Books have been written on how to manage all this data.  Whatever the approach, there are a couple of key things that are worth remembering:

·         Always understand what the question is that you need to answer – capacity reporting isn’t an end in itself

·         Don’t waste time on irrelevant data – if you can’t see how something is going to be used, consider

o   Only capturing data that you know you need to use

o   Capturing all data you think you might need but only collecting/processing/reporting on data  you actually need now

One thought might be to take the popular approach to desk/office management.  If you have something on your desk that you haven’t referred to for six months or whatever time period you consider appropriate, get rid of it. 

For Capacity Management, if you have data metrics stored that you haven’t looked at, stop processing those metrics into reports.  You immediately save the time, effort and cost associated with processing those metrics and maintaining reports based on them.  If you at least keep the captured data for a period without processing or reporting on it, as you can easily do with athene®,  you can always process that data if and when a need subsequently arises..

Andrew Smith

CEO