Self-Healing on the Cloud: State-of-the-Art and Future Challenges

Despite the enhancement on software hardness induced by development-time automatic testing and debugging tools, it remains practically impossible to create fault-free applications. The research on self-healing systems emerged to enable applications to cope with unexpected events at run-time in order to maximize their availability, survivability, maintainability, and reliability, while minimizing human intervention. With the increase in software´s complexity, in part triggered by the advent of the Internet and Cloud Computing, self-healing properties in applications are becoming a necessity. The main focus of this position paper is on presenting the issues that render unusable or ineffective the usage of the development-time Spectrum-based Fault Localization(SFL) algorithm in run-time environments. We concluded that, despite the issues found, it should be possible to devise an SFL algorithm for run-time environments that can also achieve the good results yielded by SFL at development-time.