Network Coding Based On Sparse Matrix In Distributed Secure Storage System

Network coding is a new research area in information theory, recently. Based on theknowledge of graph theory and finite-field arithmetic, this thesis briefs the principle,advantage and application of the network coding and decoding. In the distributed securestorage system, the concrete coding process of random linear network coding wasdiscussedcombining with the concept of sparse matrixes in LDPC,irregular matrixes etc. Besides,coding and decoding efficiency has been improved and complexity has been lower byoptimizing the coding, constructing Sparse Checking matrix. In the process of distributionand coding，control the degree of encoding vector. A sparse generator matrix with thecomposition of no-related to the encoding vector. Mergea sparse generator matrix with avirtual matrix to make the generator matrix of the systematic structure. Directly convert tothe sparse parity check matrix. Lead sparsity to random linear network coding. The P2Pdistribution Secure storage system is presented，based on random linear network coding.based on the principles and techniques of random linear network coding and BT, thesystem is designed. And improves the policy of Piece selection. Finally, the advantages ofnetwork coding in secure storage system and the developing direction in the practicalapplication have been given.