MENU
You are working on a complex systems integration project that will soon be deployed to the production environment. This system is replacing a system that was popular with the users and had no outstanding non-functional issues. There are multiple components that interact and these have been developed by various development and testing groups including some outsourced groups. You will be leading the performance testing effort. Although you would prefer to have a dedicated test system for this effort, you will have to use the production system. You can do this testing at a low usage time, but there will be some users on the system and you will be using real data for the tests. You are now planning your performance testing. Unfortunately, there are no requirements for the performance requirements of the system. How do you determine the acceptable performance levels for the various operational profiles?
Given the absence of defined performance requirements for the new system, a practical approach is to use the performance metrics of the legacy system as a benchmark. This method is beneficial as it provides a clear, historical baseline of what users are accustomed to and accept as satisfactory performance. Benchmarking against the legacy system ensures the new system meets or exceeds the performance levels that users already find acceptable, which can facilitate smoother acceptance and transition to the new system.