A new approach for storage balancing in wireless sensor networks

Abstract

The use of mobile sinks in data collection has received much attention in recent years. Mobility offers many benefits such as reliability, lifetime enhancement, security and so on. However, it poses a set of challenges the most prominent is data loss. Data collection using a mobile sink may induce huge latencies and packet drops within the network particularly when the sink moves with a lower speed and in an unpredictable manner. Thereby, some region might be visited more than others, which causes data loss due to buffer overflow. In an attempt to overcome this issue, we propose in this paper an efficient approach that balances the network storage and prevents packet drops. Nodes having run out of storage space offload their extra data to their neighbors in a proportional manner in function of their free spaces. In case all nodes have filled up their memories and there is no possibility for data partitioning, the sink is notified urgently to move towards the crowded zone to perform data offloading before the occurrence of any packet drop. Intensive simulation reveals that our approach significantly decreases packet loss and spread the load within the network.