Towards the Automated Engineering of Dependable Adaptive Services

The dependability of runtime composition of Cloud-Based Services is affected by many issues including scalability and reliability, which are commonly proposed to be solved by the use of adaptive software solutions. Usually, however, the use of various software paradigms to achieve such adaptability requires that the design of the formal operational model to be implicitly linked with the end implementation. Such rigid reliance upon a static model thereby constrains the adaptability and flexibility of the composition. This paper contends that in order for adaptable Cloud-Based Services to be produced, a formal model should be specified and enacted independently from the runtime implementation. The paper shows by use of situation calculus and software representation techniques for service composition, such as an Intention Description Language, how these issues can be addressed for assured runtime adaptable, deliberative systems. In conclusion, by use of a representative example and a case study, the methodology proposed is evaluated to highlight the advantages discussed, and to identify problems still to be resolved.