You have no items in your shopping cart.

Free Returns for 30 Days

Worldwide Shipping

Dedicated Customer Care

Safe & Secure Payments

Challenges in Server Capacity Planning

Server capacity planning challenges us to determine what will happen in the future, and as with any other attempt at prophecy, the chance of being wrong is high. Strange stuff happens, and it happens regularly.

On the brighter side of prognostication is that experience counts. The more historical data you collect from your servers, the better the chances are that trends can be identified and extrapolated. Even with the added complexities of virtualization across the board (server, storage, and network), historical data can lead to better guesses as to when your hardware infrastructure will need expansion and/or refreshing with more capable technologies.

Ideally, this would all be automated, but automation itself depends upon a fairly high amount of predictability. Automating manufacturing assembly lines is far easier than automating capacity planning. Where assembly lines have easily identifiable workloads, most business applications do not. Take for instance application response time across the Internet. Is slow response time due to network traffic or overstressed servers? Perhaps I/O spikes have something to do with it. Maybe memory usage expands into paging, which increases I/O. The application could be bumping heads with other applications that are also stressed, a scenario that can easily happen in a virtual server environment. The point is that applying human minds to the problem is still a better approach than depending upon AI exclusively to answer the questions.

One suggestion that is often thrown out in a cavalier manner is to know your business. You should know what the peak demands are on your computing infrastructure, when they happen, and what to do for accommodating the spikes. Even before the complexities of web applications, this was a tall order. Today it is nearly impossible to identify peak usage times, since the demand can increase at any moment and in unforeseen ways from any parts of the world.

A common answer to this problem is to install extra capacity, which might never be used or could be too little too late. A better approach is to tie a future expansion/upgrade into the extra capacity for ensuring that it will indeed be used at some point in the future. So how much extra capacity should you have? That depends on your particular situation, but common amounts vary from 15-25 percent.

Capacity planning has always been a guessing game, and some capacity planners are better at this than others. Better tools are being developed that meld capacity planning with performance monitoring, since both disciplines entail the same source data much of the time. Capacity planning is different in that the predictions are supposed to cover longer intervals of months or years, as opposed to immediate tuning efforts. Nevertheless, human minds will likely always be needed to ponder what the data really mean.

If you need further help with your capacity planning efforts, remember that the knowledgeable staff at eServer Supplies will be happy to review your situation and offer reliable advice.