Sukumar Ghosh, Ph.D.
Areas of Interest
The deployment of distributed systems in real applications has significantly increased. With the growth of embedded systems, the man-to-processor ratio is decreasing at an alarming rate. Large distributed systems view failures and perturbations as events and not catastrophic exceptions. Due to the dwindling man-to-processor ratio, it is not always feasible to ask for external intervention every time a failure or a perturbation occurs: future systems should be able to recover on their own. There are different paradigms addressing such issues: These are known as self-stabilization, self-healing, self-reconfiguration, autonomic computing, recovery-oriented computing, adaptive distributed systems etc. These topics define the primary focus of my research.
Spontaneous recovery and adaptation to changing environments are sometimes accompanied by harmful side effects. For example, in self-stabilizing systems, even a single transient failure can corrupt the entire network before recovery begins. Paths to recovery can also compromise with the safety requirements. Examples are abundant in sensor networks and P2P networks.
My research deals with various techniques for handling failures and recovery. The current areas of investigation are:
- Self-stabilizaton and games
- Adaptive and self-optimizing P2P networks
- Algorithms for sensor networks