Moe369 wrote:Reading some of your recent descriptions of your PerformanceTest services the big question I have is, what is the real difference between a "Browser User" and a "virtual user"?
Won't a VU with captured HTTP traffic produce the same effect at the server as a fully browser based playback?
--Moe369
Because eValid is a complete brower that emulates a real user -- to the point where the server cannot distinguish that it is an automated browser -- we use the phrase "Browser User" or BU to describe eValid playing back a recorded session.
If you only capture HTTP traffic, then you have a "Virtual User" or VU and in some cases, yes, a VU and and BU are the same. This will be true if the session you are handling is "navigational" and does not involve any state modifying actions at the browser level.
But this is still a little bit different than a full BU because the browser does the page component recovery in multiple threads -- it does these in parallel -- and the VU only does the retrievals in series. That can be a 10% to 80% performance difference.
If there are state-dependent actions in which the browser remembers things as it does with JavaScript) and/or if the browser interacts with the server dynamically, then the simple HTTP sequence won't accurately reproduce user behavior. That's why a BU is needed, for most heavy-JavaScript applications and for most AJAX applications. You need the complete realism of a BU -- which is what eValid delivers.
Hope this answers your question.
The eValid Team
