Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Nemo: A Low-Write-Amplification Cache for Tiny Objects on Log-Structured Flash Devices

Xufeng Yang, Tingting Tan, Jingxin Hu, Congming Gao, Mingyang Liu, Tianyang Jiang, Jian Chen, Linbo Long, Yina Lv, Jiwu Shu

TL;DR

Nemo is proposed, which enhances set-associative cache design by increasing hash collision probability to improve set fill rate, thereby reducing application-level write amplification and achieves three key objectives for flash cache: low write amplification, high memory efficiency, and low miss ratio.

Abstract

Modern storage systems predominantly use flash-based SSDs as a cache layer due to their favorable performance and cost efficiency. However, in tiny-object workloads, existing flash cache designs still suffer from high write amplification. Even when deploying advanced log-structured flash devices (e.g., Zoned Namespace SSDs and Flexible Data Placement SSDs) with low device-level write amplification, application-level write amplification still dominates. This work proposes Nemo, which enhances set-associative cache design by increasing hash collision probability to improve set fill rate, thereby reducing application-level write amplification. To satisfy caching requirements, including high memory efficiency and low miss ratio, we introduce a bloom filter-based indexing mechanism that significantly reduces memory overhead, and adopt a hybrid hotness tracking to achieve low miss ratio without losing memory efficiency. Experimental results show that Nemo simultaneously achieves three key objectives for flash cache: low write amplification, high memory efficiency, and low miss ratio.

Nemo: A Low-Write-Amplification Cache for Tiny Objects on Log-Structured Flash Devices

TL;DR

Nemo is proposed, which enhances set-associative cache design by increasing hash collision probability to improve set fill rate, thereby reducing application-level write amplification and achieves three key objectives for flash cache: low write amplification, high memory efficiency, and low miss ratio.

Abstract

Modern storage systems predominantly use flash-based SSDs as a cache layer due to their favorable performance and cost efficiency. However, in tiny-object workloads, existing flash cache designs still suffer from high write amplification. Even when deploying advanced log-structured flash devices (e.g., Zoned Namespace SSDs and Flexible Data Placement SSDs) with low device-level write amplification, application-level write amplification still dominates. This work proposes Nemo, which enhances set-associative cache design by increasing hash collision probability to improve set fill rate, thereby reducing application-level write amplification. To satisfy caching requirements, including high memory efficiency and low miss ratio, we introduce a bloom filter-based indexing mechanism that significantly reduces memory overhead, and adopt a hybrid hotness tracking to achieve low miss ratio without losing memory efficiency. Experimental results show that Nemo simultaneously achieves three key objectives for flash cache: low write amplification, high memory efficiency, and low miss ratio.
Paper Structure (36 sections, 12 equations, 19 figures, 6 tables)

This paper contains 36 sections, 12 equations, 19 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Tiny objects in KV cache
  4. The design criteria of flash cache
  5. Limitations of Existing Approaches
  6. Motivation: Miles to Go in WA Reduction
  7. Breakdown of the WA in hierarchical cache
  8. Quantifying the Impact of L2SWA
  9. Analysis of L2SWA(P)
  10. Analysis of L2SWA(A)
  11. Analysis of p
  12. Summary:
  13. Nemo Design
  14. Overview
  15. Preparing a "Perfect" Set-Group (C1)
  16. ...and 21 more sections

Figures (19)

  • Figure 1: Application-level write amplification comparison. "S" denotes Set, "SG" denotes Set-Group.
  • Figure 2: Hierarchical cache.
  • Figure 3: Breakdown of write amplification in a hierarchical cache. "P0$\sim$P1" denotes 2 flash pages in the physical HLog. "S5$\sim$S6" appears twice, but there is only one physical instance, to illustrate the differing garbage collection triggering behaviors of FairyWREN (Case 3.2) and Kangaroo (Case 3.1).
  • Figure 4: Passive object migration.
  • Figure 5: CDF of two migrations.
  • ...and 14 more figures