Distributed Storage Based On Fountain Codes And The Minimum Spanning Tree

Wireless Sensor Networks(WSN)is a distributed sensor network formed by self-organization through wireless communication.When wireless sensors in networks work in harmful or rapidly changing environments,they are more likely to suffer node damage or node energy depletion.When a sensor node fails,it loses its perception and all its stored data.This situation will seriously affect the network's ability to perceive and store data,resulting in its reduced practicability.In wireless sensor networks,the way of data storage has an extremely important impact on the persistence and integrity of data in the network.Therefore,how to improve the validity and reliability of data storage and monitoring in wireless sensor networks has become a research hotspot.At present,many scholars have studied the data storage problem in wireless sensor networks,and introduced the idea of fountain code: in each node,the XOR sum of several random source data packets can be stored.In the data recovery stage,only a part of the storage content of the nodes can be used to recover all the source data packets.However,due to the characteristics of fountain code,when the number of encoding data packets collected in the data recovery process is insufficient,only a small number of source data packets can be successfully recovered,and the efficiency of data recovery is low.In addition,most of the existing fountain code-based distributed storage algorithms rely on random walk method to disseminate sensing data in the network.When the number of hops of random walk is small,the data stored between nodes is less correlated,which will lead to the source data can not be recovered completely.When the number of hops of random walk is large,the number of data sent by the node will be greatly increased,which will lead to the rapid energy consumption and failure of the node.Aiming at the two main problems of distributed data storage in wireless sensor networks,this paper makes the following two research works:(1)Aiming at the problem of low decoding efficiency of fountain codes in the process of data recovery in distributed storage,a degree distribution optimization algorithm of LT Codes is proposed: based on the existing degree distribution,some improvements are made,then several degree distribution functions are weighted together,and the weight factor is optimized,so that the optimized degree distribution has better decoding performance.Compared with the original degree distribution,the improved method proposed in this paper can decode successfully with fewer encoding data packets collected,and improve the efficiency of data recovery.(2)In view of the contradiction between data reliability and network lifetime in the existing distributed storage algorithm based on fountain code and relying on random walk in wireless sensor networks,the concept of minimum spanning tree is introduced.Combining the improved degree distribution and minimum spanning tree proposed in this paper,a distributed storage algorithm based on minimum spanning tree and improved LT code is designed.Compared with the original algorithm,it improves the recoverability of coded data,saves the energy consumption of nodes and improves the decoding efficiency.In wireless sensor networks,especially when nodes are in fragile environments,such as harsh disaster scenarios,the reliability of data can not be guaranteed.In order to improve the reliability of important data in the network and prolong the working time of the network,it is particularly important to study the mechanism of improving the reliability of data storage and reducing the energy consumption of nodes.