Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

Making Real File Systems Faster with Applied Computational Storage - SDC 2022

SNIAVideo via YouTube

Overview

Explore a 43-minute conference talk from the Storage Developer Conference 2022 that delves into enhancing real file systems through applied computational storage. Learn how the Accelerated Box of Flash (ABOF) project, a collaboration between Eideticom, Nvidia, Aeon, SK hynix, and LANL, revolutionizes network-attached flash storage systems by implementing computational elements near storage devices. Discover the technical implementation of accelerating ZFS file systems, understand how computational storage disrupts traditional architectures, and gain insights into co-designing file systems with storage hardware. Through detailed presentations by experts from Los Alamos National Laboratory and Eideticom, examine the challenges of server memory bandwidth, data management tasks like compression and deduplication, and the evolution of NVMe Computational Storage standards. Gain practical knowledge about distributed analytical functions, performance optimization, and the open-source artifacts developed during the ABOF prototype creation. Prior understanding of NVMe, computational storage, and disaggregated storage concepts will enhance comprehension of this technical discussion.

Syllabus

Intro
Acknowledgements
Traditional HPC Storage
All Flash File Systems
Quick Aside HPC Storage Pipeline
File System Services Limitation
ABOF - Hardware Overview
Computational Storage Can Help!
Integrating ABOF
ZFS Interface for Accelerators (Z.I.A)
Data Processing Unit Services Module (DPU-SVC)
Data Movement - Zoff Alloc Buffer
Theory of Operations
Performance Analysis
Conclusions and Future

Taught by

SNIAVideo

Reviews

Start your review of Making Real File Systems Faster with Applied Computational Storage - SDC 2022

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.