A Light-Weight Hot Data Identification Scheme via Grouping-based LRU Lists

摘要

Real-world workloads generally exhibit high skewness in access patterns, and it is a consensus that separating hot and cold data may greatly improve storage system performance such as Solid State Drive (SSD) garbage collection (GC) performance. To achieve this, the key issue is how to accurately identify hot data, which is really challenging due to the large diversity and dynamics of workloads. In this paper, we propose a light-weight and high-accuracy identification scheme, which is developed via a group of Least Recently Used (LRU) lists and requires only a small amount of memory and CPU cycles. We further deploy our scheme on SSDs with DiskSim simulator, and results show that comparing to two state-of-the-art identification schemes, our scheme further reduces SSD GC cost by up to 59.1% (62.1%), and saves 44.3% (77.5%) of computational cost. Due to the light-weight and parameter-insensitive feature, our scheme can be easily deployed at various system levels and adaptable to different workloads.