Monday, February 16, 2009

IT Metrics That Matter to IT and Data Center Professionals

Written by Tsvetanka Stoyanova

IT metrics is a vast topic and there are many methodologies used in theory and fewer of them in practice. IT metrics lie on the crossroad of business and technology. IT metrics measure different aspects of the activity of a company and there are different sets of metrics for different companies.

The need to measure the performance of a data center is obvious. If you can't measure something, you can't manage it. IT metrics provide feedback about the performance of a data center and based on this feedback managerial decisions are made.

What Is the Reason Behind IT Metrics?

The idea of measuring performance is not new to technical people. However, unlike metrics in computing, which though not necessarily precise, manages to capture easily quantifiable values – i.e. downtime measures the time a network is down, or bandwidth measures the width of a channel, IT metrics which are used in business are a bit elusive and subjective. Actually, if there is something more elusive than an IT standard, it must be IT metrics.

Needless to say, when the conclusions reached via applying IT metrics are untrue, this is misleading and IT metrics become of less use. What is more, if the result is very untrue (or fundamentally wrong), this can lead to making the wrong decision and in this case IT metrics are not only useless but they become harmful. Just imagine that you decide to measure downtime based on the number of computers you have – i.e. if you had 100 computers, this would lead to100% uptime. But unfortunately you have only 90 computers, so you can never reach 99.99%, not to mention 100% uptime.

While this example might be pretty lame and extreme, it is possible for a manager to apply IT metrics in that way. IT metrics when misused can be disastrous.

Sample IT Metrics for Use in Data Centers

When used properly, IT metrics can be of great use. The hard part is how to use IT metrics to their advantage.

The first point that needs to be cleared is what a company will measure. Only then comes the question how to do it and which set of IT metrics to choose. According to Forrester analysts, “The key to success is choosing a small number of metrics that are relevant to the business and have the most impact on business outcomes. The five metrics that meet the criteria for relevance and impact are investment alignment to business strategy, business value of IT investments, IT budget balance, service level excellence, and operational excellence. These five metrics should form the core of an IT performance scorecard. ”

The inappropriate choice of IT metrics methodology is the first common problem. You choose a very clever-looking methodology, which looks like the answer to your prayers but it turns out that even if the methodology is not full of technical errors (i.e. measure uptime based on the number of computers), it doesn't measure what you want to measure simply because it is not designed for it.

Another common problem at the stage of choosing an IT metrics methodology is that managers “get greedy” - i.e. to want to have the most comprehensive and complete IT metrics system, which will measure everything and everybody. While such a system might be possible in theory, in practice this approach fails. If you look at the quotation above, Forrester analysts stress that a small number of relevant metrics is what works best. And this is the case no matter what you try to measure.

It is possible to use simultaneously many sets of IT metrics and to measure many aspects of the activity of your data center. For instance, you can measure performance in general, or how green your data center is. One of the best IT metrics set to measure performance in general is the Key Performance Indicators (KPIs) methodology. If you want to learn more about it, the A Hierarchy of Metrics article will give you a basic idea upon which you can expand further.
Green IT metrics are also popular, especially for data centers. There are also sets of IT metrics to measure mainframe performance. Metrics of financial performance are traditional. Actually, there are hundreds of sets and methodologies to measure everything – one piece at a time. Certainly, there is no lack of IT metrics for those who are eager to use them.

IT metrics come in all shapes and sizes. Many organizations, research institutes or even separate companies and consultants develop their own sets of IT metrics. It is impossible to say which one of the many sets is best because even if the methodology is perfect it may not meet your exact needs.

It is not possible to recommend one and only IT metric set which works always and for everybody. The catch is that a good IT metric set should be universal, yet tailored to the needs of the particular company. Only if these conditions are met, will you receive results that can be trusted and used as a basis for decisions.


Recent Posts