Explore a 25-minute conference talk from USENIX ATC '20 that delves into AC-Key, an innovative Adaptive Caching system for LSM-based Key-Value Stores. Learn about the challenges of read amplification in LSM-tree structures and how AC-Key addresses these issues through efficient caching algorithms. Discover the three main caching components managed by AC-Key: key-value cache, key-pointer cache, and block cache. Understand how the system adapts to workload changes and utilizes a novel caching efficiency factor to optimize performance. Examine the implementation process using RocksDB modifications and review evaluation results comparing AC-Key's performance to RocksDB and offline fix-sized caching schemes in various workload scenarios.
AC-Key - Adaptive Caching for LSM-based Key-Value Stores
Overview
Syllabus
Introduction
Internal Operations
Hierarchical Adaptive Caching
Challenges
Implementation
Phase Change Test
Summary
Taught by
USENIX
Related Courses
-
MatrixKV - Reducing Write Stalls and Write Amplification in LSM-tree Based KV Stores
-
Calcspar - A Contract-Aware LSM Store for Cloud Storage with Low Latency Spikes
-
SplinterDB - Closing the Bandwidth Gap for NVMe Key-Value Stores
-
PinK - High-speed In-storage Key-value Store with Bounded Tails
-
Improving Performance of Flash Based Key-Value Stores Using Storage Class Memory as a Volatile Memory Extension
-
FPGA-Accelerated Compactions for LSM Based Key-Value Store
