Exploration of Adaptation Space - Linking with Efforts in Service-Oriented Computing
There have been a lot of efforts on self-adaptation in the area of
service-oriented computing. Systems are composed by combining existing
services, and adapt typically by selecting and switching from alternative
services. General problems of service composition select a service for each
involved task for given global goals (functions, quality (non-functional)
constraints, and optimization criteria).
Our work extends this setting to select a set of candidate services for each
task, to avoid greedy non-optimal adaptation and to reduce runtime overhead for
adaptation. As a technical approach, our work uses graphs to model
slightly-different functions of similar services. This approach allows to
efficiently identify alternative services as well as to see the potential
trade-offs between strength of goals and availability of alternatives. Although
this work targeted the communities of world wide web and service-oriented
computing, it shows an instance study about how to identify and explore
adaptation space (inside solution space) at development-time and at runtime.