分布式存储系统中混合数据布局算法

Hybrid Data Placement Algorithm in Distributed Storage System

数据布局算法是分布式存储系统的基础性算法,也是提高数据处理效率的关键。针对节点负载和通信延迟等存储节点状态,提出了一种衡量存储节点可用性的通用方法,并在分析了已有算法的基础上,综合各种算法的优点,提出了一种混合数据布局算法。该算法根据存储节点可用性不同而采取不同的数据冗余策略。通过对比分析,证实该算法在存储量与通信量方面具有较大的优越性。

Data placement algorithm is one of the basic algorithms and the key to enhance data processing efficiency in distributed storage systems.This paper presented a universal approach to measure storage node availability based on state variables such as load of storage node and network delay.After analysis the existing data processing algorithms,we raised a new hybrid data placement algorithm.Different redundancy policies are applied to storage node with different node availability.According to our analysis,this algorithm has its advantages in store and communication costs.