Skimming Of Test-Case Execution Bandwidth On A Balanced Cloud Service Through An Worker Role With Logistic Regression Determination

Cloud environments pose new challenges to the traditional testing techniques by means of architectural challenges, effective utilization of bandwidth available and security. Effective bandwidth utilization is regarded as the most important challenge to the testers while designing the test cases that are meant to run on the live environment since unconcentrated test case design yields in more compute hours and plays a role in QoS reduction to the real customers for cross bandwidth utilization. Hence an attempt has been made in this paper to skim the execution bandwidth utilized on a cloud service that is running across multiple servers across the globe by imparting the logistic regression method. The tests are run against the cloud services hosted on windows azure cloud platform and bandwidth utilized with regards to compute hours are measured against the bandwidth used on normal case without the worker role involving in the bandwidth reduction and about 52% reduction overall is achieved with 93% of confidence level.