一种自适应负载的I/O调度算法

I/O调度算法对磁盘阵列(RAID)性能具有至关重要的影响。虽然已有很多典型的I/O调度算法在一定负载情况下可获得较好的性能,但很难有哪一种算法在各种负载情况下均能获得很好的性能。本文提出了一种智能RAID控制模型,结合C4.5决策树和AdaBoost算法实现负载自动分类,根据负载变化和性能反馈情况动态调整I/O调度策略,实现面向应用需求的自治调度。模拟实验结果表明,自适应调度算法具有较好的适应性,在各种负载情况下优于现有的I/O调度算法,尤其适用于多线程混合负载环境的I/O性能优化。

An Adaptive I/O Scheduling Algorithm

The I/O scheduling algorithm has crucial influence upon the disk array (RAID) performance. Although there are many typical I/O scheduling algorithms which get preferable performance in certain workload cases, unfortunately, it is difficult for a single universal scheduler to be capable of providing superior performance across all system workloads. This paper introduces an intelligent RAID control model and a new method to resolve this problem. Combining the C4. 5 decision tree and the AdaBoost algorithm to automatically recognize the type of the workload, adjusting the I/O scheduling strategy dynamically according to the workload changes and performance feedback, and a self-optimizing scheduler is implemented to accommodate the demand of applications. The simulation results show that the adaptive scheduling algorithm can adapt to a wide variety of workloads. It outperforms the existing I/O scheduling algorithms under various workloads,and is especially suitable for I/O performance optimization in the environment of multi-threaded and mixed workloads.