A novel approach to energy efficient fault tolerant time synchronization protocols for wireless sensor networks

Due to recent innovations in the area of wireless technologies, it is feasible to deploy inexpensive tiny disposable and low power devices throughout a wide area for the measurement of physical parameters like temperature, pressure, humidity etc. These small devices are known as sensor nodes (SNs) and their deployment is termed as wireless sensor network (WSN). SNs have resource constraints in terms of storage capacity, battery power and processing capabilities. WSN was originally developed for military applications, but later it has found its use in lots of public domain. WSNs involve monitoring events and communicating sensor readings to some designated nodes in the network known as sinks. Each node uses a local clock to timestamp an event at this node and there is not global clock as such. However, local clocks may differ from each other due to variable drift over a period of time. Hence, WSN applications require partial or full time synchronization and it is important that message exchanged by the SNs must be time stamped by each sensor's local clock. Time synchronization in WSNs is also important because time based queries can be answered only if all the nodes have a common notion of time or are synchronized within a permissible limit. Further, in WSNs, the clocks do not remain well synchronized over a long period without periodic resynchronization. Therefore, to maintain the notion of global time, the clocks of the nodes must be resynchronized periodically and in a manner that consumes minimum possible energy. In this paper, we have proposed an Energy Efficient Fault Tolerant Time Synchronization (EEFTTS) Protocol for WSNs which efficiently synchronizes the WSN with minimal energy consumption. Unlike many synchronization protocols, where pairwise message exchanges take place, proposed scheme synchronizes in steps at different topological layers with minimal message exchanges. EEFTTS also detects certain node failures and corrects it by selecting alternate node as new synchronizing entity i.e cluster head Simulation results show that proposed synchronization scheme provides low power consumption as compared to existing algorithms like RBS and TPSN.