Good question. The goal of the integration is to get end user experience measurement from real devices while the server is loaded.
1. The load itself is generated at protocol level by NeoLoad with a regular Load Test
2. The goal of the perfecto integration is to execute regularly Perfecto scripts on a limited number of real devices in parallel of the load testing. This regular execution at device level allows you to get trends as the server is loaded.
As load testing results and Perfecto results ends in NeoLoad results, you can : 1. Get the real device metrics according to the load 2. correlate all kind of measurements to analyze the server and device behaviors. (ex: What are my real device metrics when the server CPU is 90% ? What is the impact of server side performance degradation on the client side?)