A Blackboard-based Decision Support Framework for Testing Client/Server Applications

Abstract

To assist a solution to the problem of the test environment spanning multiple platforms, this paper proposes a decision support framework with the blackboard model to integrate all complementary features into a single automated test environment for multi-platform client/server applications. Before testing client/server applications, the input into this framework are testing tools with different approaches and client sites which are going to run the test. The planning agent will make a decision dynamically and produce a testing plan to allocate testing tasks to these testing tools to client sites. Two complementary features for testing client/server applications are illustrated in this paper to demonstrate how the framework works. The concept of mobile agents is applied to launch the test driver to different client sites, execute the tests and bring back the test results from client sites as well as the trace file from the server site for inspecting the interaction behavior among clients. Based on the multicast framework, the same test data can be broadcasted to multiple clients sites to run the tests simultaneously and the test results can be returned from client sites for examining the problem of repeated executions.