Try   HackMD

EPF Update 2

During the past two weeks I focused on MEV-boost and its implementation. To boil it down, MEV-boost is simply a communications software that sits right outside of the evm (which comes with some advantages like supporting client diversity). Right now MEV-boost only allows for full block building, so proposers (assuming they aren't also a block builder) cannot capture MEV.

Now this implementation has a few weaknesses, and in the short term, I think the biggest weakness is that relayers are very trusted.

Thoughts from the EPF

I got in contact with echno from the cohort who is also working on PBS, who will probably focus on removing trust from the relay. His second update referenced EigenLayer's implementation of MEV-Boost and his take on it, which I found was a great project. I then read Barnabe's feedback to echno's update which had alot of great insight that further scope out some small challenges in PBS/ PBA implementations. Anti-spam mechanics via RLN's, and building a "synthetic" Eigenlayer with a different stake were areas that seemed interesting and I will for sure look into soon after writing this update.

EigenLayer + MEV-Boost implementation

EigenLayer's implementation of MEV is kind of a derivative of a partial block auction. The biggest difference is the proposer builds an alternative block B_alt as well as the Proposer_part of the block (can be thought of as appended to the Builder part).

Anyways, this implementation eliminates the liveness risk that comes with the trusting of relayers from MEV-boost since the block proposer can just release B_alt as the next block.

Closing thoughts

Although I still don't have clarity on what exactly I want to implement, I am almost certain I want to do something involving PBA's, and am getting close to scoping out the exact problem I want to focus on. I will look further into RLN's and look for other PBA implementations like EigenLayer's for inspiration.

Next Steps

Scope for the problem I am solving for.
Write a project proposal.

Some sources from this update:

https://hackmd.io/@layr/SkBRqvdC5
https://ethresear.ch/t/how-much-can-we-constrain-builders-without-bringing-back-heavy-burdens-to-proposers/13808/7
https://hackmd.io/@echno/S1QbuQaHs
https://www.youtube.com/watch?v=PS9SlHGJlrU&t=1556s
https://efdn.notion.site/Review-echno-s-update-on-mev-boost-relays-7c37138dd94144829700d581ab80cdbd

Last changed by 
 
ballestar
0
121

Read more

EPF Update 1

This is my first in a series of weekly updates as a part of the Ethereum Protocol Fellowship (Cohort Three), since these are weekly, I'll keep them short and sweet. My first week was devoted to exploring possible research topics related to MEV. I went into this week with a pretty rudimentary understanding of MEV, but I just enough to know I wanted to dig deeper into the rabbit hole. I spent this week reading material from flashbots and Alex Stokes to get a base level understanding of the future of MEV, particularly In Protocol Proposer Builder Separation (IPPBS) and the implications of builder centralization. I was pretty much trying to get a better understanding of the problem. TLDR: delegating block-building to an off chain "builder market" could cause a nasty flywheel that will inevitably weaken Ethereum's censorship-resistant properties. There's a set of probabilistic outcomes if we continue with the current state of PBS. A few of these outcomes are: 1. Since the builder market inherintly has economies of scale, they will become more powerful overtime, centralize, and gain in-protocol influence by owning validators (from their builder profits) 2. Builder centralization will lead to censorship at the builder-level and possibly the network level. In the short term, IPPBS is good for the dominant builder in the short term, but will fragilize the network in the long term. Oookay, so MEV and PBS is clearly an important problem to solve. What does the thinking around these problems look like right now? I've skimmed over a few popular solutions proposed by Alex Stokes, Vitalik , and Barnabe. Right now the leading design is Two Slot PBS, but there are ideas like PEPC, and ideas like Partial Block Auctions that I need to dive deeper into the future.

Nov 17, 2022
Read more from ballestar
Published on HackMD