Traditional load testing typically means testing back-end systems. Simply put, you simulate a request and measure the time between receiving the first and the last byte. While the concept is simple, any performance engineer will tell you that it is not necessarily easy.
We are dramatically oversimplifying what it takes to load test because we want to take a look at the next step – browser render time. For years I thought browser render time was important but not vital. In fact, I used to believe that I wanted to test only the elements that have control over such as the application, the servers that host the application, and the network on which those servers reside. I can’t be held accountable for how fast or slow someone’s computer or device is. It’s not my fault if they have 58 tabs open!
Then one day, like a Mack truck it hit me. Like scales falling from eyes I saw that user experience is paramount. I realized that I needed to refocus my efforts and knew that all my effort was really about delivering the best user experience. The user doesn’t care that the back-end is performing well. He only cares about his experience. With the advent of Single Page Applications, render time became even more critical.
In this discussion, I’ll cover ways by which we, as performance engineers, can incorporate real browsers into our process and how Total Performance Consulting has incorporated hybrid testing, combining both traditional protocol based performance testing with real browsers, to measure and analyze. We’ll discuss issues we have discovered that would only have been found using a hybrid approach as well as how to implement a hybrid approach without breaking your budget.