盘阵中基于平衡超图划分的自适应请求 并发与负载平衡策略

作为一种并行系统,盘阵性能极大地依赖于设备间负载平衡;为了减少执行时间,单个I/O请求需要通过多个设备并发完成,负载平衡并不意味着请求并发,而请求并发也不能保证盘阵负载平衡.因此,必须将二者结合起来,才能使盘阵性能得到更好的优化.为此本文提出基于平衡超图划分的自适应数据分布策略,并提出了两个目标函数同时进行请求并发与负载平衡的控制,根据I/O访问模式优化盘阵性能.根据两种I/O benchmark合成负载进行模拟试验表明,该策略比传统的单方面自适应负载平衡策略或基于超图的请求并发策略效果都有一定优化.

An Adaptive Strategy Based on Balanced Hypergraph-Partition for Both Load Balance and Intra-request Concurrency in Disk Array System

Disk array is a high performance storage system for exploiting intra-request concurrence and inter-request parallelism.Load balance among disks is necessary for better inter-request parallelism,and an I/O request should be served by multiple disks for intra-request concurrence.This can be achieved for optimal data placement.But file access pattern is changing.System should redistribute data blocks adaptively.In this paper,we present a strategy of balanced hypergraph-based adaptive data placement,and give two objective functions for the optimization of both load balance and intra-request concurrence.We have done lots of simulations to evaluate our method and compare it with traditional optimization strategies such as adaptive load-balance and hyper-graph placement for intra-request concurrency.The result manifests that our strategy is a useful one.