An approach for selecting best available services through a new method of decomposing QoS constraints

Quality of Service (QoS) plays as a discriminating factor for selecting appropriate services that meet the given user’s non-functional requirements during service composition. There is a compelling need to select suitable services quickly so that the composition can meet dynamic needs. Recently, local selection approaches for QoS-based selection have been put forward toward reduced time complexity. A methodology for selecting the best available service combination for a given user requirement (workflow) with a new method of decomposing QoS constraints is proposed in this paper. The methodology consists of two phases, namely ‘Constraint Decomposition Phase’ and ‘Service Selection Phase’. In the Constraint Decomposition Phase, a unique method is proposed to decompose the given non-functional (global or workflow level) constraints into local constraints for individual tasks in the workflow. Each individual task with its local constraints forms a subproblem. In the Service Selection phase, each subproblem is resolved by finding the best available service from its respective service class using an iterative searching procedure. A prototype has been implemented, and the low computation time of the proposed method makes it well suited to dynamic composition. The proposed method of decomposing constraints is independent of number of services in a service class, and the method is applicable to any combinational workflow with AND, OR and Loop patterns. Further, a new method for computing response time of OR execution pattern which guarantees successful execution of each path in an OR pattern is a remarkable contribution of this work.