You might think that capacity planning is something that only large Information Systems organizations need to do, but actually all of us change our work habits over time as we acquire new software. It can be fascinating to watch the computer system become taxed over time. But this fascination has its practical side—if we watch closely enough, we'll know exactly what to do to improve the performance of the system as the demand for it increases.
In this chapter, we'll give you some tips on how to stay one step ahead of the demand for your system, whether it's a network server or a desktop computer. Capacity planning begins with keeping records of the performance of your system over time. These records can become so huge as to be practically useless if you're not careful, so a significant part of capacity planning is thoughtful and organized record keeping. Once you have good records to sift through, you can get to the analysis of those records.
The analysis is really just the application of the concepts of bottleneck detection we have explained in Chapters 3 through 7, from the perspective of watching how your computer usage habits have changed in the past, and where they are heading. Mercifully, we will not repeat those chapters here.
As a fringe benefit, the whole task of bottleneck detection is greatly simplified with even a little capacity planning. It's easier to see what's changed than to start from scratch to determine what's wrong. That Memory: Non-Paged Pool Bytes, didn't that used to be lower? Was it that new application we got, or was it adding TCP/IP to the network protocols? Just a little history is worth its weight in charts here.