Flash-Conscious Cache Population for Enterprise Database Workloads

Host-side flash caching has lately emerged as a suitable and e↵ective means of accelerating enterprise workloads. However, cache management for flash-based caching is di↵erent from traditional DRAM-based caching. A flash cache sits underneath the DRAM cache. Its position in the hierarchy combined with the unique characteristics of flash, calls for a di↵erent cache management solution. Specifically, cache population, an aspect of cache management which is not attributed much importance in DRAM caches, becomes crucial in flash-based caches. In this paper, we first present a performance evaluation of three popular open-source flash cache implementations: flashcache, bcache, and EnhanceIO . We evaluate them under an industry-standard database benchmark and identify their limitations. We demonstrate that several shortcomings are due to sub-optimal cache population. We propose a novel set of techniques for cache population, and present the design of the Scalable Cache Engine (SCE) – a new flash cache solution that incorporates our cache population techniques. We demonstrate that SCE remarkably outperforms the existing open-source solutions: 45% higher throughput, 55% lower latency, 12⇥ faster cache warm-up than flashcache, and 95% less memory usage than EnhanceIO .