A Comparative Case Study on the Engineering of Self-Testable Autonomic Software

A survey on the landscape of self-adaptive systems identified testing and assurance as one of the most neglected areas in the engineering of autonomic software. However, since the structure and behavior of autonomic software can vary during its execution, runtime testing is critical to ensure that faults are not introduced into the system as a result of dynamic adaptation. Some researchers have developed approaches and supporting designs for integrating runtime testing into the workflow of autonomic software. In this paper, we describe a comparative case study performed on three autonomic applications that were engineered to include an implicit self-test characteristic. The findings of our study provide evidentiary insight into the benefits and software engineering challenges associated with developing these types of systems.