Research On Storage Overflow Problem Of Nodes In Wireless Sensor Networks

In recent years,wireless sensor network technology has been developed rapidly.It has been widely applied to various fields,especially it provides the possibility for monitoring data under the harsh environment such as volcano and ocean.Due to these special environment,it is not suitable to install long-term fixed base stations,and data items collected by sensor nodes therefore must be stored inside the network.After a period of time,it leads to storage overflow problem as nodes with higher frequency of data collection collect more data items.Storage overflow problem seriously affects the normal work of the sensor network and it has two levels: partial storage overflow and overall storage overflow.For partial storage overflow,the total amount of overflow data is less than or equal to the entire available storage space of the sensor network,while overall storage overflow problem is on the opposite.Storage overflow problem in wireless sensor network is relative new research.And existing research focus on total energy consumption of the network,ignoring that different residual energy of node,which can easily cause the node to die as the energy consumption of a node is high.Thus this paper studies partial and overall storage overflow problem with the goal of maximizing the minimum residual energy of sensor nodes.For partial storage overflow problem,we proposed a solution which is based upon data redistribution.And we also demonstrate that data redistribution problem for maximizing the minimum residual energy of data nodes is NP-hard.Data redistribution is redistributing overflow data to nodes which have available storage space.Based on the solution,we proposed a data redistribution algorithm which is based upon maximum flow of network.The algorithm can solve partial storage overflow problem optimally,but its time complexity is relatively high.In order to meet the practical application of the sensor networks,we proposed a BFS-based data redistribution algorithm which has low time complexity and can effectively balance energy consumption between nodes.For overall storage overflow problem,we proposed a solution which is based upon data aggregation.Data aggregation can eliminate redundant data between overflow data,so that overall storage overflow problem can be turned into partial storage overflow problem,and then it can be solved by data redistribution.Based on the solution,we proposed a degree-constrained data aggregation algorithm.And the algorithm outperforms than existing algorithms in term of total energy consumption and maximizing the minimum residual energy of sensor nodes.In order to achieve better performance,we further put forward a data transfer algorithm.Data transfer algorithm which integrates data aggregation and redistribution can store some overflow data in the process of data aggregation.And it can further reduce energy consumption of nodes and extend the lifetime of the sensor network.