Using blackboard system to automate and optimize workflow orchestrations

Abstract

Automated composition and optimization of workflows in Service Oriented Computing (SOC) is a challenging research area. However, there are numerous problems yet to be completely resolved, such as: How to automate the solution in a service oriented environment with redundant services competing with each other? How to formally quantify the performance metrics based on user requirements both at the task level and at the overall workflow level? How to select the optimal set of services in order to maximize the user happiness? How to cope with the dynamic behaviors such as user changing his requirements and failing resources? We present a smart approach based on the Blackboard Systems for the generation of workflows in service enriched environments such as the Grid and the Cloud Computing. We devise a mathematical model to automate the blackboard based workflow generation. Our model defines a ”happiness measure”, that, on the basis of user specified requirements and priorities for various activities in the abstract workflow, identifies optimal possibilities of service orchestrations. Then we discuss an example scenario to demonstrate these formal techniques and show a speedup analysis of our parallelized Branch and Bound algorithm based implementation. We also discuss possibilities of other more efficient algorithms.