QZFS: QAT Accelerated Compression in File System for Application Agnostic and Cost Efficient Data Storage
Talkat USENIX ATC '19 · on Wednesday, 2019-07-10· at 15:20· USENIX ATC Track I: Grand Ballroom I–VI·
summary
Integration of Intel® QAT ASIC into ZFS accelerates gzip (de)compression at the filesystem layer, adds a compression service engine that selects hardware or software per-block by compressibility and source size, and implements a vectored I/O QAT offloading module that reconstructs pages into contiguous buffers for zero memory copy. The prototype reports up to 5x write throughput on FIO and more than 6x cost-efficiency for genomic pipelines, with selective offload avoiding small-block and low-compression waste.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:32Background
- 0:52NVMe SSDs
- 1:30Cost of ownership
- 2:59Overview
- 3:54Design Considerations
- 5:02Architecture
- 6:40Data Reading
- 8:27Loading Module
- 9:59Evaluation
- 10:52File Macro
- 12:28Postprocessing
- 13:01Application Layer
- 14:44FIO Benchmark
- 15:43ORS Approach
- 16:52QAT Latency
