SRS: Scalable and Robust Snapshot Algorithm For Wireless Sensor Networks

Global state determination of distributed systems gives information that is useful in applications of checkpointing, distributed garbage collection, deadlock detection, distributed algorithm termination detection, event boundary detection, distributed debugging, etc. Determination of the global state in a wireless sensor network needs a different approach than of traditional distributed networks due to unique properties of WSN like a constraint on resources, energy constraints, remote hazardous environments where sensors are deployed, network topology that is dynamic in nature, etc. In this paper, we propose a distributed snapshot algorithm that is scalable, robust, and provides efficient communication among WSN nodes which is based on tree architecture. We designed an algorithm in such a way that it can easily be scaled to any number of nodes without affecting the performance of the underlying system and it can effectively handle packet loss so that the communication cost of the whole system is maintained as minimal. We performed simulations in OMNET++ and did the analysis of results based on the number of messages exchanged, response time, and performance in packet loss situations.