Dynamic Context-Aware Service Composition, Brokerage and Monitoring in Service-Oriented Computing

点击翻译按钮获取中文摘要

Service composition has long been an active research domain that enables a service broker to generate a composite service adapted to the complex needs of their clients. Well-known individual problems such as constraint satisfaction, context-awareness, data integration, and data privacy have all found various independent solutions in the literature. However, a broader look at the situation reveals that these problems are in fact interwoven and their interaction creates other problems that individual solutions cannot solve. Our solution proposes a sound theoretical basis to an integrated solution to all these problems as a whole. An article from the Financial Post reported in March 2017 that the Government of Canada is potentially losing up to 1.3B$ yearly in taxation revenue due to international e-commerce failing to charge taxes. This problem stems from the fact that service providers involved in international e-commerce cannot be expected to be aware, nor have authority to apply all international constraints/policies/regulations that apply to the specific context of every customer to which they provide a service. Furthermore, e-commerce has a tendency to create more business volume to larger companies that can afford investments in online service collaboration with other companies. Automated service composition augmented with context/constraint-awareness can help more specialized and commercially unconnected service providers to reach for their clientele. For example, a transaction with a local service may prove to be more appropriate compared with a transaction with a foreign service provider after international policies are applied to a transaction. Having an integrated solution for constraint/policy/context management would thus potentially foster the economic viability of contextually specialized (e.g. local) service providers.The problems stated above are all related, from various perspectives, to the automated manipulation of an integrated data model over the composite service. Despite this common denominator, these problems have not met with an appropriately integrated solution. More specifically, various solutions were provided to the problem of constraint-awareness in composite services. However, they all consider constraints as being inherent to the whole composite service without consideration which part of the composite service they may apply. Also, none of the existing solutions provide an integrated solution considering different sources of constraints, especially externally-defined constraints. This project aims at the development of a new integrated constraint/contextual-aware model for the service composition process, as well as for the execution of composite services. This model is based on the notion of possible-worlds semantics, where the computational context and the constraints are modeled over a multidimensional context space.