DYSARM

Dynamic Service Adaptation using Model-Driven Engineering

A Smart City will adapt the services it provides to the needs of its citizens, and to conserve and preserve its resources. Such adaptation must occur in real-time, and may not be anticipated and designed for in advance. New software models are required to support reasoning about the run-time behaviour of the City’s services, assist in the adaptation decision-making process, and affect the required adaptation in real-time. Run-time models that reflect the state of execution of running systems enable the specification of required adaptations. These models can reduce the time taken to comprehend system behaviour, identify problems, propose changes and verify the results of proposed changes.

DYSARM will develop a runtime model environment and toolkit that will improve the understandability and, consequently, the reliability of adaptations in dynamic service-oriented computing applications through support of, at the model level: adaptation specification; monitoring and validation of the executing system; and automated adaptation of services at runtime. With models providing views of the executing system, its impact on the behaviour of the city structures and services and its continued adaptation, the designer can reason over the semantics and current state of the system, and also the adaptation, at an appropriate level of abstraction.

This project’s objectives are:

to provide for models as a view of the run-time behaviour of a city’s services;
to support software services adaptation by using runtime models to provide a richer semantic base for decision-making;
and to develop a set of tools that support reasoning about runtime adaptation and executable service-oriented systems at a higher level of abstraction, reducing the time taken to understand system behaviour, identify problems and verifying the results of proposed changes.

Task 1 – Maintain Runtime Model of City-Wide Software Services

This research will investigate city-wide service composition concerns, with a view to maintaining the “most appropriate” runtime information to support dynamic adaptation of those services. Links between the dynamic state and system constraints will be maintained, as will the causal relationship between runtime system and design models. The runtime system will be monitored for “relevant” changes that may trigger adaptation, and notify an adaptation decision-making process of the required adaptation.

Task 2 – Dynamic Adaptation of City-Wide Software Services

A model and supporting environment for dynamically adapting city-wide software services will require definition of a metamodel for runtime adaptation specification. Adaptation requests may occur as a result of the monitoring of the city’s runtime services and environment, or from explicit requests from the software architect. This task aims to apply model-driven engineering techniques, combined with adaptation protocols, to effect software adaptation at runtime while maintaining acceptable quality-of-service.