Robustness Evaluation of Controllers in Self-Adaptive Software Systems

An increasingly important requirement for software-intensive systems is the ability to self-manage by adapting their structure and behavior at run-time in an autonomous way as a response to a variety of changes that may occur to the system, its environment, or its goals. In particular, self-adaptive (or autonomic) systems incorporate complex software components that act as controllers of a target system by executing actions through effectors, based on information monitored by probes. However, although these controllers are becoming critical in many application domains, it is still difficult to assess their robustness. The proposed approach for evaluating the robustness of controllers for self-adaptive software systems, is aimed at the effective identification of design faults. To achieve this objective, our proposal is based on a set of robustness tests that include the provision of mutated inputs to the interfaces between the controller and the target system (i.e., probes). The feasibility of the approach is evaluated in the context of Znn.com, a case study implemented using the Rainbow framework for architecture-based self-adaptation.