Extracting Behavioral Models from Executions Data in Web Services Environments

Abstract

We investigate the problem of extracting web ser- vices’ protocols basing on execution log events, i.e., how to recon- struct protocols’ specifications expressing the observed behaviors from service’s execution traces ?

We propose a comprehensive logic-based approach to extract business protocol models expressing the external behavior of deployed web services. Recorded execution traces are transcribed to a facts base and a set of structure inference patterns is trans- lated to a corresponding production rules base. The conceived knowledge base is explored by a reasoning engine to infer the various elements describing the finite state machine representing the target service protocol and thus, the behavior model contained in the execution traces log files is extracted.