CRDTs and the Quest for Distributed Consistency

Channel:InfoQ · Length:43:39 · Published:2018-10-02 · Views:68491

Watch on YouTube

Summary

Martin Kleppmann's QCon London 2018 talk on keeping data consistent in distributed systems without a central leader. It surveys CRDT algorithms for syncing devices and collaborative editing over peer-to-peer networks.

Why HN recommends it

Strongly recommended as a clear, accessible survey of the CRDT landscape.

I also highly recommend the talk — HN, 2019

Distributed systems · Algorithms & data structures

Subtitles (excerpt)

my name is Martin Chapman you might know me from this book that I published about a year ago called designing data intensive applications this book is not about CRD T's instead it's a very broad overview of the architecture of large-scale data systems and problems of distributed systems what I'm going to talk about today is a different area so I am a researcher at the University of Cambridge except today actually I am formerly on strike you might have you might have gathered I mean this seriously does there's a dispute going on at the moment at many UK universities where the powers that be want to essentially cut back our pension schemes in a very radical and unjustified way which amounts to a massive pay cut and really is not justified so as part of this I am on strike I'm still going to give to talk to you because our we have no problems with koukin it's our university that we have a problem with but just to say that I am giving this talk as a private individual who happens to be interested in distributed systems rather than ok so let's talk about some technical stuff so what I'd like to talk about is collaborative applications and so a classic example of this would be Google Docs so

On Hacker News

Recently going through this book. I highly recommend it, and any videos by the author, Martin Kleppmann I especially like this one: > CRDTs and the quest for distributed consistency [VIDEO] — syspec, 2020-07
I think the original statement is defensible. Many of the original OT papers proposed algorithms that have since been shown not to converge in particular cases (especially the TP2 puzzle). It took a while for the fog to clear, in part because TP2 requires concurrent edits from 3 peers. I also highly recommend the talk "CRDTs and the Quest for Distributed Consistency"[0] from Martin Kleppmann, whi… — raphlinus, 2019-06
Recently saw a fantastic intermediate talk on this subject called CRDTs and the Quest for Distributed Consistency[0]. One product of the research was automerge[1], a library for handling collaborative editing of JSON-like data models [0]: [VIDEO] [1]: [LINK] — hardwaresofton, 2018-10

Related videos