基于网络编码的P2P备份系统毕业设计（论文）.doc

NCStore: 基于网络编码的P2P备份系统 计算机科学技术学院 李钧 指导教师 陈学青 摘要：随着存储设备容量的不断提升，对重要数据进行有效备份和保护的需求也在快速增长。传统的C/S模型具有高硬件需求、高维护性需求、低带宽利用率等缺点，而P2P存储模型由于具有分布式存储的特征，可以有效利用网络中的存储及带宽资源，从而提供廉价、快速、可靠的数据存储服务。 在数据传输过程中使用网络编码，由于其在中间节点编码的特征，可以有效提高网络吞吐率和鲁棒性。因此，在存储系统中使用网络编码，一方面可以提高数据的可用性，另一方面，当系统中存在数据丢失需要进行数据修复时，网络编码可以有效降低修复过程中的带宽开销。本文分析了存储系统中使用网络编码的参数选择，以避免网络编码带来的线性相关的问题。基于分析结果，本文设计并实现了一个使用网络编码的数据备份系统NCStore。 实验结果表明, 与副本及纠删码方式相比, NCStore可以降低数据修复时的传输时间。此外，使用网络编码可以保持数据可靠性的同时，降低带宽使用。 关键词：网络编码 P2P 备份系统。 Abstract: With the increasing of the capacity of hard disk, the demand for the preservation of important data is also increasing rapidly. While the traditional C/S architecture suffers from high cost and low bandwidth usage, the peer-to-peer architecture can utilize the storage and bandwidth resource in the network and thus provide cheap, fast and reliable backup storage. By using network coding during the data transmission in the network, it may bring two main benefits: the improvement of network throughput and a high robustness. Therefore by using network coding in the storage system, it can not only provide high data reliability, but also reduce bandwidth cost during the data regeneration, when data stored in the system gets lost. In this paper, we analyze how to choose a suitable network coding scheme to avoid the problem of linear independence. Secondly, we propose a design of a real backup system and implement it. Results of experiment in LAN show, compared with replica and erasure codes, NCStore consumes less time for data regeneration. In addition, the network coding scheme can not only keep the data reliable, but save bandwidth usage as well. Keywords: network coding, P2P backup system 1 Introduction In this paper, we focus on the question that how to save important data in the network and keep it durable, or reliable with low cost. The simple principle of all the solutions is using redundancy. As the rapid increasing of hard disk capacity, it is common that there is free space in a hard