Flexible Ontology-Independent and QOS-Enabled Dynamic Web Services Composition Using Google Distance

摘要

Problems with ontology-dependent approaches for dynamic web services composition occur when ontologies are poorly maintained or unavailable in a domain, or the costs of maintenance exceed their benefits. To resolve such problems, this paper proposes the use of Google Distance as an ontology-independent method for the semantic similarity matching stage of web services discovery. Further, it provides an improved method for matching QoS parameters in the operational similarity matching stage of web services selection. The methods are embedded within a simple architecture for dynamic web services composition. We provide comparisons between our and existing QoS-enabled service discovery approaches in two major categories of web composition methods: workflow-based and AI-based. Our implementation exercise clearly shows the performance tradeoffs that help implementers to better assess an overall QoS-enabled dynamic composition method for its suitability to application-level services workload characteristics.