Towards a Decision Approach for Autonomic Systems Adaptation

The autonomic computing paradigm arises to tackle the complexity of today's dynamic heterogeneous and ubiquitous systems. The architecture of autonomic systems based on the MAPE-K loop dynamically adapts itself when the context changes. However, decisions within phases of the MAPE-K loop must be taken and therefore fit a suitable architecture. Ongoing researches take decisions upon context changes of the system and are a well studied topic in different fields, however, most of them use predefined decisions which do not cover all the contextual changes or decisions which support only a very specific type of application. In this work we propose an approach dealing with the decisions that must be taken within MAPE-K loop for system adaptation, in the context of autonomic computing. An exceptional architectural configuration of systems relies on consistent decisions taken by the analysis, planning and execution phases of the aforementioned loop. Autonomic systems undergo dynamic environment changes and therefore decisions lead the architectural configuration of the system to success or failure.