An efficient approach for storage balancing in wireless sensor networks

The use of mobile sinks in data collection has received much attention in recent years. In fact, mobility was introduced to solve problems that occur in data gathering with static sink such as hotspots, quick energy depletion of sensor nodes and so on. Using a mobile sink provides an effective mechanism to improve reliability, security as well as connectivity within the network. Nevertheless, the sink's mobility poses new challenges, especially when the sink follows an unpredictable movement while gathering data. In this case, the network will experience huge latency and suffer from significant packet loss particularly when sensor nodes do not have enough memory storage to buffer collected data between two consecutive visits of the mobile sink. In this paper we propose a new approach in which sensor nodes cooperate to manage the storage and prevent packet drops. When a node's memory is almost full, it offloads its data to its neighbor nodes in function of their free spaces. In case there are no suitable neighbor nodes with sufficient storage space, the sink is urgently notified about the overloaded region that needs to be rapidly dumped. Simulation results reveal that our proposed approach balances the storage load and decreases drastically packet loss.