Normally, we measure the performance requirements in terms of latency, which is the response time for a given transaction per user. For example, the time an e-commerce website takes to process the order, after the customer clicks on the Buy button for the checkout screen to appear, is the response time for the checkout page.

Typically, you would test this operation multiple times and note the average response time. We have to note that different operations on the e-commerce website, such as adding items to a cart, search for new products, and checkout order, might have different response time values.

In addition to response time, performance can also be based on transaction throughput, which is the number of events that take place within a given time period, usually one second.

Still, in the e-commerce example, you can get the total transaction time for the user (in this case, the transaction time to buy something in the website), by adding the time for...