Rainbow: An Approach to Facilitate Restorative Functionality Within Distributed Autonomic Systems

Core autonomic behaviours, namely self-healing, selftuning, and self-adaptation, are now prevalent features of modern software architectures. Current implementations have typically relied on static rule and goal oriented approaches that respond to operational triggers to provide self-adaptive behaviour. This paper builds on the introspective nature of the previous work of the authors, namely the Cloud framework and Neptune language, to produce Rainbow, a system that enables storage and restoration of configuration data from an autonomic system. It is shown that the exposition of such configuration data enables system evolution to be captured and documented for analysis, and later restoration, essential abilities for successful anticipation and handling of changes in the operational context of an executing system. A working implementation of the proposed system is introduced, with conclusions made based on the future directions the research introduced in this paper can take.