ACHIEVING AUTONOMIC SERVICE ORIENTED ARCHITECTURE USING CASE BASED REASONING

Service-Oriented Architecture (SOA) enables composition of large and complex
computational units out of the available atomic services. However, implementation of
SOA, for its dynamic nature, could bring about challenges in terms of service
discovery, service interaction, service composition, robustness, etc. In the near future,
SOA will often need to dynamically re-configuring and re-organizing its topologies of
interactions between the web services because of some unpredictable events, such as
crashes or network problems, which will cause service unavailability. Complexity and
dynamism of the current and future global network system require service architecture
that is capable of autonomously changing its structure and functionality to meet
dynamic changes in the requirements and environment with little human intervention.
This then needs to motivate the research described throughout this thesis.
In this thesis, the idea of introducing autonomy and adapting case-based reasoning
into SOA in order to extend the intelligence and capability of SOA is contributed and
elaborated. It is conducted by proposing architecture of an autonomic SOA
framework based on case-based reasoning and the architectural considerations of
autonomic computing paradigm. It is then followed by developing and analyzing
formal models of the proposed architecture using Petri Net. The framework is also
tested and analyzed through case studies, simulation, and prototype development. The
case studies show feasibility to employing case-based reasoning and autonomic
computing into SOA domain and the simulation results show believability that it
would increase the intelligence, capability, usability and robustness of SOA. It was
shown that SOA can be improved to cope with dynamic environment and services
unavailability by incorporating case-based reasoning and autonomic computing
paradigm to monitor and analyze events and service requests, then to plan and execute
the appropriate actions using the knowledge stored in knowledge database.