Model driven development and dynamic composition of Web services.

摘要

Web Services (WS) have emerged as a new component-based software development paradigm in a network-centric environment based on the Service-Oriented Architecture (SOA), the XML open standard description language, and HTTP transportation protocol. Legacy software systems can incorporate WS technology in order to be reused and integrated in a distributed environment across heterogeneous platforms. While WS is gaining its momentum toward wide adoption in the software industry, there are two critical issues yet to be addressed before its power is fully realized: (1) the migration of legacy distributed software systems toward WS applications and (2) the innovation of new infrastructures and languages in support of WS application development. The contribution of this dissertation is in these two directions.; First, a comprehensive, systematic, automatable, and language-neutral approach is presented toward reengineering legacy software systems to WS applications, rather than rewriting the whole legacy software system from scratch in an ad hoc, language-specific manner. It is noteworthy that this approach is not specific to reengineering WS applications but can be generalized to reengineering legacy software systems to other applications. Moreover, this approach offers a means for modeling assets exchange in both horizontal and vertical directions (along the meta-model stack).; Second, with the dynamic features of both service consumption and provisioning in distributed environments, WS applications are subject to dynamic composition. As such, in a bottom-up order, this dissertation presents an infrastructure for dynamic WS composition and its high-level programming model based on a hybrid of logic programming and imperative programming. In particular, with the logic programming paradigm and the rule inference engine support, not only is autonomous composition achieved but also WS selection specification can be seamlessly integrated with the composition process, which is necessary for achieving customizability, optimization, and Quality of Service (QoS) guarantees for dynamic composition.