Network coding overhearing management policies based on data packet switching and sorting algorithm

摘要

Network coding-based overhearing management policies can greatly improve coding ratio and system throughput of coding nodes. The existing strategies, such as FIFO, best effort service, and historical information-based etc, are all based on isolation mechanism of the node information, they do not fully consider the relation among the cached information of nodes, as a result, the node cache information is frequently replaced or deleted by mistake, affecting the coding ratio of coding nodes and the decoding ratio of destination nodes. Therefore, there is a lot of space for improvement. In “X” pattern network topology, network coding overhearing policy based on data packet switching and sorting algorithm is proposed in this article. First, when overhearing buffer is full, meanwhile, the sink codes have listened new packets, codes will drop the new packets but record the IDs of the dropped packets; Second, intermediate codes get the packets information of overhearing buffer by the means of improved Epidemic Algorithm, then sort the packets of coding queue according to the packets information of overhearing cache. Finally, sink nodes delete the packets which have been used for decoding, and add the newly-coming packets in the cache. Theoretical analysis and simulation demonstrate that the improved algorithm gets better coding ratio, higher system throughput and lower delay than the traditional overhearing policies.