"Dealing with Extreme Heterogeneity and Dynamicity in Next Generation Middleware"

This research project is jointly carried out by University College London's Dept. of Computer Science (Software Systems Engineering Group), and by Lancaster University's Computing Department (Next Generation Middleware Group). This project aims at dealing with extreme heterogeineity and dynamicity in next generation middleware.

With the success of the Grid, there is an accelerating trend towards diversity in terms of both application domains and also, crucially, in the underlying networked infrastructures in use. Furthermore, there is an ever-increasing dynamicity in the Grid which present-day static architectures cannot cope with. The premise of this project is that the present Globus and web services-based Grid architecture is unable to cope with these pressures and that a paradigm shift in Grid middleware architecture will be required in the medium term. We will therefore investigate a radically new Grid middleware architecture that is inherently extensible, and is based on a fine-grained, open, component-based structure that is highly configurable and self-adaptive. Furthermore, we will leverage Model Driven Architecture principles to automatically generate appropriate ‘families’ of component configurations for divergent infrastructures (e.g. ad-hoc environments, local clusters, the global Internet), together with bridges that enable these families to interoperate. Crucially, our MDA tool will also generate runtime models that manage the generated configurations according to autonomic computing principles.

To evaluate our architecture, which builds on substantial complementary principles developed by the partners, and, indeed on a substantial existing software and infrastructure base, we will port and extend existing e-Science applications developed in the RealityGrid project. We will also focus on how the key non-functional properties of security and dependability can be handled in the architecture.