Capacity Planning - What Is It?
The term capacity planning is used as a way to describe set of heuristic as well as formal techniques that have been made to figure out the type and number of servers needed to host given set of workloads. It's certainly not an immediate priorities. This is perceived most of the time as time consuming and expensive as a result, it is something that does not deserve much attention.
Not only that, IT departments usually perceives planning as responsibility of server providers. What is meant by this, having transferred the responsibility of capacity planning to vendor and they are no longer have interest in it.
In comparison to other fields of engineering, IT engineering is mostly dominated by the idea rule of thumb or the "good enough" concept. You wouldn't like to fly on a plane designed with the rule of thumbs but, it's accepted largely in big IT project to become grounded on this rule. This is just one of the main causes of high failure rates for multiple IT projects. Included in these is the preparation of need report.
Keep in mind that good enough is never enough and we have to explore much better ways of designing IT projects and among the areas is the capacity planning. We have proved that using optimization algorithms rather than rule of thumb firms can acquire big savings. It's a bit obvious that the optimization algorithm is providing more effective solutions than the rule of thumb. What appears to be less obvious is that, you can get it done efficiently and that you're making big savings. So far as savings are concerned, we have to explore different proficient and effective approaches that are listed below.
Data collection - everyone knows that the quality of analysis hinges onto the quality of data being used by the analysis. IT departments should be aware how difficult it is to get data collection that is accurate. And problems that customers not having accurate and updated list of servers, clients might not have accurate list of the security credentials needed to access servers and so forth are some of the common issues. Solving the problems here is the crux that is requiring cooperation from clients. If you want to learn more about capacity planning, you can visit https://en.wikipedia.org/wiki/Capacity_planning.
Parallelism - in gathering workload and inventory data, there have been many ERP software and tools that are tested but most have failed. In order to get the data as soon as possible, separate thread is very important to collect data from every server.
Data analysis - for projects with more than 100 servers, this is undoubtedly a time consuming task. Normalization of workload data is among the important activities that makes it possible to compare the processor use in various servers.