40x less overhead! Rateless Invertible Bloom Filters

40x less overhead! Rateless Invertible Bloom Filters - Part 3 of 3

Рет қаралды 759

Күн бұрын

Practical Rateless Set Reconciliation by Lei Yang, Yossi Gilad, Mohammad Alizadeh: arxiv.org/abs/2402.02668
Part 3 in a 3 part series on bloom filters
Chapters:
00:00 Nerd Shade
01:36 Set Reconciliation Overview
03:08 vs Invertible Bloom Filters
05:04 Design Goals
06:40 Ratelessness
08:42 Probability Mapping
12:20 Building a RIBLT
15:18 Reconciliation using RIBLT
19:30 Conclusion

Пікірлер: 17

@LeoDaLionEdits 2 сағат бұрын

I never even heard of a bloom filter before watching this series, but after I feel enlighted. Thank you for being so enthusiastic and teaching this in an easy way to understand

@chopper3lw 4 күн бұрын

This is so farkin amazing, thanks for spending the time to cover it.

@gulamahsan5902 7 күн бұрын

You are amazing .. I wish I had a computer science teacher like you in my academia.

@n0computer 7 күн бұрын

Thanks so much! You just made our day ☺️

@manan6247 18 күн бұрын

Great videos, I was in third year of undergrad when I first discovered your channel, I was looking for a research topic and your videos inspired me a lot. Thank you

@n0computer 18 күн бұрын

Thank you so much!

@kickeddroid 17 күн бұрын

Very Juicy indeed!!! Love your videos man

@kickeddroid 17 күн бұрын

Something that interests me about probabilistic data structures. Is how they can be used to measure causality between order of events. If it’s a probability and it’s eventually consistent then couldn’t we achieve a rate less version of Order Reconciling. I know this wasn’t the point of the paper but lovely to ponder.

@atom-06 17 күн бұрын

Love the videos

@atom-06 2 күн бұрын

can u do the signal protocol?

@n0computer Күн бұрын

Oooooh that’s a great idea! We have a full slate for the next few months, but might just start reading docs…

@atom-06 15 сағат бұрын

@@n0computer thannk you

@triforce42 23 сағат бұрын

To me, "Zeroth" is confusing in this context. I also haven't heard "Oneth." I understand it means "First" and "second" respectively but if it's confusing generally, maybe that's something worth changing in the future.

@n0computer 18 сағат бұрын

Totally agreed, and thanks for the feedback! We’ll skip it moving forward ☺️

@dipi71 13 күн бұрын

Doesn't rsync already have this kind of improvement? (If not, the rsync project ought to publish something about their algorithm.) Cheers!

@n0computer 12 күн бұрын

It’s been a while since we looked, but rsync famously uses a rolling hash function for file reconciliation, which is integrated into the wire protocol. Would be worth investigating!