ethrd

Hey Farcaster ethr&d friends, I wanted to share a quick summary of our latest beam call

This one focused on the peer-to-peer networking layer

Read on 👇 & watch the recording attached

Coming back to the core topic of this call: p2p-networking. As a quick introduction:

One of the main objectives is to come up with an effective networking protocol for peer discovery and message passing (gossiping) between nodes. Such protocol must strike the right balance between latency, message transportation resilience, and bandwidth usage.

A key parameter is the size of messages distributed among peers. The substantial increase in size of post-quantum signatures in the context of the beam chain compared to current signatures necessitates exploring more optimal networking protocols

Setting the stage, @justindrake laid out a strawman proposal, leveraging the existing beacon chain network framework (libp2p) and adding a two-level permissionless aggregation layer - utilizing recursive aggregation for post-quantum signatures first on a subnet-level and later on a global level

@agemanning from SigmaPrime continued, highlighting a possible alternative to the currently used gossipsub protocol (for message gossiping) by proposing a more modular version called generalized gossipsub. This lets a network choose exactly how information spreads within it, even changing the rules for different types of information or at different times - improving its reliability and efficiency

@agemanning from SigmaPrime continued, highlighting a possible alternative to the currently used gossipsub protocol (for message gossiping) by proposing a more modular version called generalized gossipsub. This lets a network choose exactly how information spreads within it, even changing the rules for different types of information or at different times - improving its reliability and efficiency

https://github.com/libp2p/specs/pull/664

Pop from the EF p2p networking team followed up with a different alternative (gossipsub v2), representing an evolution of the currently used gossipsub protocol, aiming at minimizing the duplication of messages floating across the network

Pop from the EF p2p networking team followed up with a different alternative (gossipsub v2), representing an evolution of the currently used gossipsub protocol, aiming at minimizing the duplication of messages floating across the network

https://ethresear.ch/t/doubling-the-blob-count-with-gossipsub-v2-0/21893

Next up Kamil from Quadrivium presented a network topology for faster message distribution among validators called GRID, and shared their experience using it in the context of building a client on the polkadot network

Next up Kamil from Quadrivium presented a network topology for faster message distribution among validators called GRID, and shared their experience using it in the context of building a client on the polkadot network

https://hackmd.io/@kamilsa/B1gqs7wqye

Lei from MegaETH afterwards showcased Set Reconciliation as a technique to minimise redundant messages. It allows nodes to efficiently find the differences in their data sets and exchange only the missing information.

Lei from MegaETH afterwards showcased Set Reconciliation as a technique to minimise redundant messages. It allows nodes to efficiently find the differences in their data sets and exchange only the missing information.

https://dl.acm.org/doi/pdf/10.1145/3651890.3672219

Closing the call, Mihir from PierTwo presented their work on libp2p written in C, followed by Bing from the zen-eth team with their implementation of lib2pb written in zig

p.s.

As a quick follow-up on the previous beam call around post-quantum cryptography, Emile presented his findings on shorter aggregate signatures: his work on hash-based SNARKs using the proof system WHIR (over FRI) shows top of the class results with regard to proof size as well as codebase size for prover and verifier

Feel free to check out all beam call related content in our PM repo here:

p.s.

As a quick follow-up on the previous beam call around post-quantum cryptography, Emile presented his findings on shorter aggregate signatures: his work on hash-based SNARKs using the proof system WHIR (over FRI) shows top of the class results with regard to proof size as well as codebase size for prover and verifier

Feel free to check out all beam call related content in our PM repo here:

https://github.com/ethereum/pm/tree/master/Breakout-Room-Meetings/beam-chain

Although the beam chain is still just a proposal for a future version of Ethereum’s consensus layer, we were pleased to host more than eighty curious people on the call. We are grateful for the growing interest in the project.

Feel free to DM me if you also wanted to join one of our upcoming calls on ‘minslack’ (a more optimal validator exit queue design), Attester-Proposer-Separation and Three-Slot-Finality.