MOLAR: A cost-efficient, high-performance SSD-based hybrid storage cache

Yi Liu, Xiongzi Ge, Xiaoxia Huang, David H Du

Research output: Contribution to journalArticlepeer-review

Abstract

This paper proposes a deMOtion-based, fLash-awARe hybrid storage cache model, named MOLAR, to effectively integrate Flash-based Solid-State Disks (SSDs) into traditional dynamic random access memory (DRAM)-based memory storage systems where SSDs serve as the Tier-2 cache, while DRAM is considered as the Tier-1 cache. We found that conventional cache algorithms designed for DRAM perform poorly in SSDs due to the limited write cycles and asymmetric read/write performance of Flash memory. In MOLAR, a Flash-aware I/O path structure is designed to adapt the asymmetric read and write performance of SSDs and, moreover, to reduce useless write operations. A new control metric, demotion count, is validated to wisely select the evicted blocks from DRAM to reside in the SSD. Besides, for SSD can improve internal data placement from data access hints, the Logical Block Addresses in the SSD are grouped into the long-lived region and the short-lived region self-adaptively via a heuristic control algorithm based on the change of the block demotion count. Through trace-driven simulations, the overall hit ratio in MOLAR outperforms two traditional policies by 1.44-5.34%. The average access latency in SSDs is reduced by 3.5× to 4.5×. Moreover, write amplification is effectively reduced by ∼36% in two typical Flash address-mapping policies.

Original languageEnglish (US)
Pages (from-to)2061-2078
Number of pages18
JournalComputer Journal
Volume58
Issue number9
DOIs
StatePublished - Sep 29 2014

Keywords

  • access hints
  • demotion count
  • hybrid storage cache
  • solid-state drives

Fingerprint Dive into the research topics of 'MOLAR: A cost-efficient, high-performance SSD-based hybrid storage cache'. Together they form a unique fingerprint.

Cite this