Analyzing TCP/IP Performance

Performance problems occur when a system and its network are not operating as effectively as they should as indicated by slow response times and a decrease in users’ productivity. These problems can be intermittent or can indicate a growing strain pointing to capacity issues. Causes can be multifaceted and include both hardware and software origins or can be quickly solved with proper configuration settings.

Because of the multi-faceted nature of TCP/IP issues, this section serves as an analysis tool by providing a checklist of possible factors. By properly analyzing and defining the problem in terms of a set of symptoms and potential causes, you can either solve the performance problem yourself or can provide support with documentation.

When you analyze performance, you must also consider your organization’s priorities, such as the following goals:

  • To maximize throughput in order to achieve the maximum data transfer rate
  • To “fill the pipe” and run at full capacity

As you test and fine-tune settings to address specific factors in a problem area, record the results and note any unusual interactions or behavior. You may want to change company standards and create a checklist to accommodate new procedures or settings.