Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HACache: Leveraging Read Performance with Cache in a Heterogeneous Array

Jialin Liu, Liang Shi, Dingcui Yu

Abstract

In cost-sensitive deployments, RAID arrays may combine SSDs with different performance levels. Such heterogeneity arises when aging SSDs degrade yet remain usable, or when failed drives are replaced with new devices of explicitly better performance. While this reduces procurement cost, it creates performance challenges: traditional striping mecahnism distributes requests evenly, but slower SSDs become bottlenecks, leaving faster ones underutilized and limiting overall bandwidth to the slowest drive. To address this, we propose HACache (Heterogeneity Adaptive Cache) for read-intensive workloads. HACache introduces high-performance SSDs as read caches to rebalance request distribution. First, we formalize the request diversion problem and solve it formally. Second, to support optimal diversion ratios searching at runtime, HACache adopts a two-phase request diversion ratio adjustment mechanism. Finally, a cache capacity regulation is adopted to adapt quotas for each backend SSD based on hit rates and request diversion needs. This design maximizes bandwidth utilization. Experiments show HACache improves heterogeneous RAID read performance significantly, with bandwidth gains of about 35\% in typical mixed configurations.

HACache: Leveraging Read Performance with Cache in a Heterogeneous Array

Abstract

In cost-sensitive deployments, RAID arrays may combine SSDs with different performance levels. Such heterogeneity arises when aging SSDs degrade yet remain usable, or when failed drives are replaced with new devices of explicitly better performance. While this reduces procurement cost, it creates performance challenges: traditional striping mecahnism distributes requests evenly, but slower SSDs become bottlenecks, leaving faster ones underutilized and limiting overall bandwidth to the slowest drive. To address this, we propose HACache (Heterogeneity Adaptive Cache) for read-intensive workloads. HACache introduces high-performance SSDs as read caches to rebalance request distribution. First, we formalize the request diversion problem and solve it formally. Second, to support optimal diversion ratios searching at runtime, HACache adopts a two-phase request diversion ratio adjustment mechanism. Finally, a cache capacity regulation is adopted to adapt quotas for each backend SSD based on hit rates and request diversion needs. This design maximizes bandwidth utilization. Experiments show HACache improves heterogeneous RAID read performance significantly, with bandwidth gains of about 35\% in typical mixed configurations.

Paper Structure

This paper contains 28 sections, 1 equation, 16 figures, 1 table, 3 algorithms.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Heterogeneity in Large-Capacity SSD Arrays
  4. Bandwidth Underutilization in Heterogeneous Arrays
  5. Phenomenon Description
  6. Cause Analysis
  7. Opportunities and Challenges of Caching Techniques
  8. HACache: Adaptive Cache Design for Heterogeneous RAID
  9. Overview
  10. Optimization Objective of Cache Diversion Ratios
  11. Formal Problem Definition
  12. Problem Solving
  13. Dynamic Adjustment of Cache Diversion Ratios
  14. Limitations of the Optimal Cache Diversion Ratio Algorithm
  15. Two-Phase Dynamic Valve Value Adjustment Strategy
  16. ...and 13 more sections

Figures (16)

  • Figure 1: Measured vs. theoretical bandwidth of 4‑drive heterogeneous RAID
  • Figure 2: Proportion of requests queued at each SSD under 4KB random read workload
  • Figure 3: Proportion of requests handled by each SSD in the queue under 4KB random read workload
  • Figure 4: Bandwidth of each SSD under 4KB random read workload
  • Figure 5: Comparison of cache schemes for heterogeneous RAID
  • ...and 11 more figures