Conflict-free replicated datatypes solve distributed data consistency challenges

ably.com

61 points·stichers··3 comments

If you were wondering why there's another post about CRDTs here on hacker news, so soon after this one (https://news.ycombinator.com/item?id=28998767) I can answer that.

As I was finishing up the blog post above, I started speculating on whether HN was interested in the topic. And what kind of feedback it would get, if so.

My experiment with Nikita's list of CRDTs resources was instructive, to say the least, particularly as I'm something of a noob here.

So this time I've made sure that:

- the post works on mobile (well, as much as I can be sure) - I've spelled out the TLA (FLA?) in the title - the article is more than a list. It's much more.

I look forward to the discussion this time!

I did a bit of work on the topic (which remains preliminary), trying to reinvent the wheel a little bit and also implementing a naive (but correct) LWW-graph structure, and I doing some light thinking around applications [1]. There is one potential application of conflict-free distributed graphs like this that would interest me to further study: distributed SLAM in robotic swarms. I would be happy to pair with someone with access and know-how around the right simulation environments to work together.

[1] https://github.com/agravier/crdt-study

https://github.com/ipfs/go-ds-crdt

https://news.ycombinator.com/item?id=29028787

Reply on news.ycombinator.com