Maximizing I/O Throughput and Minimizing Performance Variation via Reinforcement Learning Based I/O Merging for SSDs

Merging technique is widely adopted by I/O schedulers to maximize system I/O throughput. However, I/O merging could increase the latency of individual I/O, thus incurring prolonged I/O latencies and enlarged performance variations. Even with better system throughput, higher worst-case latency experienced by some requests could block the SSD storage system, which violates the QoS (Quality of Service) requirement. In order to improve QoS performance while providing higher I/O throughput, this paper proposes a reinforcement learning based I/O merging approach. Through learning the characteristic of various I/O patterns, the proposed approach makes merging decisions adaptively based on different I/O workloads. Evaluation results show that the proposed scheme is capable of reducing the standard deviation of I/O latency by 19.1 percent on average, worst-case latency by 7.3-60.9 percent at the 99.9th percentile compared with the latest I/O merging scheme, while maximizing system throughput.

1 Introduction

Solid state drives (SSDs) are now widely deployed in computing systems. They differ from traditional hard disks on various aspects, such as high random access performance, low power consumption, and erase-before-write. However, traditional I/O stack designs are mainly towards HDD performance enhancements, paying little attention to taking advantage of SSD features [10], [13], [16], [19]. I/O merging adopted by I/O schedulers is an example of such designs [31], [32].