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.