Today on Datanauts, we look under the hood of distributed storage systems to explore erasure coding. Is erasure coding just a fancy name for RAID? What happens if we lose a storage node and have to recreate the missing pieces? Holy east-west traffic, Batman! It’s erasure coding today on the Datanauts podcast.
Our guest is J Metz, Research & Development Engineer for Advanced Storage at Cisco Systems. You can follow J on Twitter and read his blog, which includes technical content, political musings, and lots of pictures of Jeeps, at jmetz.com.
The Datanauts and Dr. J discuss the general concept of distributed storage, offer common examples, and explain the similarities and differences between distributed storage and RAID.
They drill into how erasure coding works and then explore issues around bottlenecks, performance, and repair.
Go full storage nerd with this show, and then check out the links with additional information. The links are posted just below our sponsor message.
Sponsor: FutureNet
VMware’s FutureNet is a networking-focused, invitation-only event being held during VMworld this August. You’ll hear from industry leaders and expert practitioners about new and emerging technologies that will transform the network. Request your invitation at vmware.com/futurenet.
Show Links:
Basics
Everything You Wanted To Know About Storage But Were Too Proud To Ask: Part Chartreuse – Webinar
Erasure Coding Wiki
Storage Performance Benchmarking: Introduction and Concepts – Webinar
Intermediate
Modern Erasure Codes for Distributed Storage Systems. Presentation at the 2016 Storage Developer Conference – PDF
Erasure Codes for Large-Scale Distributed Storage – YouTube (oldie but a goodie)
Advanced
Network Coding for Distributed Storage Systems – PDF