Recent Discussion of Invertible Bloom Lookup Tables at Monster Scale Summit 2026

Hello everyone,

There was recently a talk on IBLTs at Monster Scale Summit 2026 by Pet Helland. I noticed they seemed to be an improvement on Merkle Trees for full table anti entropy repair. I tried building a baseline use-case for it using text-based WALs but I suspect the reasoning logic may be different at a larger scale. What would be your opinion of it? Is it something you are looking into doing?