Research On The Data Storage Scheme In Wireless Sensor Networks

The main task of wireless sensor network(WSN) is to efficiently provide users with accurate and valuable environmental information. WSNs are data-centric wireless networks in which users focus on how to effectively transfer and store the data collected by sensors to ensure efficient, reliable and timely access rather than hardware equipment and application scenarios. Therefore, data storage in WSNs is an important issue. Based on different storage policies, existing methods of data storage in WSNs can be divided into three types including external, local and data-centric storage. In this thesis, our work mainly focuses on data-centric storage.The existing data-centric storage schemes made an ideal assumption on the node distribution which cannot reflect the actual distribution, which causes unbalance between nodes’ storage load and energy consumption. To solve the problem, we propose a data storage and routing algorithm based on non-uniform distribution so that event data can be stored uniformly in WSNs according to the distribution of nodes. The main idea of this algorithm is to estimate the actual distribution of nodes in the network and disperse sensed data through the hash function and the rejection method. The rejection method is applied in routing to avoid duplicate routing paths, thereby balancing the energy consumption of nodes. Meanwhile, we regard the spatial distance among nodes as the similarity of the data stored in the local storage, and establish the search program.Our proposed algorithm can achieve more uniform energy consumption, more efficient search service and a longer lifetime of networks.In this thesis, we also investigate the problem of low efficiency in data storage space. In WSNs, nodes may produce data periodically. Therefore, processing the data with delay may result in memory overload and data loss in storage nodes or even cause extra energy consumption for redundant routing. In order to reduce data redundancy in local storage, we propose a data storage algorithm based on Bloom filter(BF), and thealgorithm can provide energy efficient storage by reducing redundancy. Our proposed algorithm can deal with redundant data by BF while some errors are allowed and it can reduce energy consumption and save storage space. According to the bit-array table, we build up similarity search, and it can provide timely responses for queries.