Handle returns to executing client

[mnxfst]

The ptest-server is expected to run in a distributed environment having n parallel ptest-server instances. Since the client needs to know about the (accumulated) results each run produces, there must be a way to return those information to the client.

Since the execution is triggered using a HTTP server call, the results could be included in the response. Unfortunately this works only for a small number of test plan recurrences. Therefore either the client must provide a callback URL the server must be aware or the client needs to constantly poll results from different source (eg. database).

Another way to handle this could be a central ptest-server instance or another dedicated control instance which is ordered to execute a distributed test case, colllects all results, stores them using a unique session identifier and provides the data on a different HTTP call back to the caller. In that case the caller could constantly pull for results and must not provide a callback url.

[mnxfst]

The client knows when results should be available and requests the values after a defined time-period if there are no results, the client sleeps and comes back later