行星际互联网中基于Raptor码的分布式存储策略

摘要

针对行星际互联网节点存储资源严重受限及存储可靠性问题，提出一种基于Raptor码的分布式存储策略（DSSR）。采用定向随机漫步机制，使得源数据包能够迅速遍历整个网络。在预编码阶段利用基于ID的估计方法进行网络全局信息估计，无需额外的全局信息估计阶段。通过设计合理的数据包选择机制使得最终编码度分布趋于理想度分布。分析和仿真结果表明，与相关文献提出的分布式存储策略相比，该方法大幅度减少了数据包传输时的随机漫步步长，同时提高了译码性能，简单易行。%To solve the limited storage resource and the data storage reliability problem of Interplanetary Internet,a novel distributed storage strategy based on Raptor codes (DSSR)is proposed.According to the proposed strategy,source data packets are quickly transmitted to every node in the cluster by use of a directional random walk mechanism.The ID-based estimation method is used to estimate the global information at the pre-coding phase,the values are obtained without additional global information estimation phase.The reasonable procedure of XORing packets is designed so that the distribution of code degree tends to the ideal distribution.As presented by the analyses and simulations,random walk steps is greatly reduced compared with a relevant distributed storage strategy,while improving the decoding performance.