Flashbots Transparency Report: MEV-Share, SGX block building & ETHDenver

Overview

The past few weeks have been quite eventful with ETHDenver, market volatility, and various announcements. This edition will cover development, research and events from February 10th until March 16th. Due to the sheer amount of content since the previous report, this edition will be the longest yet. I’ve done my best to keep things succinct and engaging. Enjoy!

MEV-Share: programmably private orderflow to share MEV with users

Just in time for Valentine’s Day, @bert introduced MEV-Share: a permissionless and private matchmaking protocol between users and searchers. MEV-Share builds on MEV-Boost by further unbundling the transaction supply chain, enabling collaboration between searchers and users.

Through MEV-Share users will be able to receive MEV back by revealing parts of their transactions to searchers through programmable privacy. Searchers will have access to this additional information and bid in an auction for the right to execute against it.

MEV-Share explores how programmable privacy, preference expression, and credible commitments can enable decentralized block building and can be seen as a proto-SUAVE that can progressively evolve into a fully decentralized network.

We’ve released the client library for the MEV-Share Matchmaker, featuring examples of programmable privacy and using the MEV-Share endpoint to backrun transactions.

Block Building inside SGX

In the previous Transparency report I highlighted the work of @metachris and Frieder Paape of running a prototype of Geth inside SGX and syncing with Ethereum mainnet.

Fast forward to earlier this month and they have now successfully implemented a fully functional block builder inside SGX. This development marks a significant step towards decentralized block building and reducing the risks posed by exclusive order flow.

Running a block builder inside an SGX enclave allows block builders to create economically efficient blocks without knowledge of the contents of the user’s transactions.

On March 3rd when the article was published, the SGX builder was live on the Sepolia testnet, however as of publishing this report, the SGX builder has begun landing blocks on mainnet! Blocks built by the Flashbots SGX builder can be identified by the extra_data of the block: Illuminate Dmocrtz Dstrib Prtct.


From Etherscan
The Github repository has documentation and an example for running Geth in Gramine, a libOS for SGX enclaves. Join the conversation in the forum post!

MEV-Boost

All-time high for daily MEV payments to Block Proposers

The market turmoil on March 11th triggered a surge in on-chain activity and MEV opportunities, resulting in a new daily all-time high for validator payments via MEV-Boost. The total payments made to block proposers amounted to 7691 ETH, nearly twice the previous record set during the FTX collapse. @bert published a thread looking into the activity that unfolded and the impact it had on both the Relay and Builder market of MEV-Boost.


From mevboost.pics

This record breaking amount of ETH being extracted by searchers resulted in many negative externalities and outcomes for users, the most notable example being the bot that made over $2M from a user that failed to set their slippage limit on a trade, as described by @bert and separately by KyberSwap.

Diversification in the relay market

A considerable diversification can be seen in the relay market with Agnostic Relay and Ultra Sound Relay gaining market share in February and March. In November 2022 the blocks relayed through the Flashbots Relay peaked at ~60-70% of mainnet blocks. That number is now about 25%.


From mevboost.pics

This diversification in the relay market has contributed to a noticeable decrease in the percentage of proposed blocks sourced from non content-agnostic relays. The previous report indicated that 58% of blocks were sourced from these relays, this figure has now dropped to 37%


From mevwatch.info

MEV-Boost community calls

@ralexstokes is facilitating MEV-Boost community calls to discuss MEV-Boost and the Relay and Builder implementations. Two calls have been held so far with great turnout and valuable conversations.

MEV-Boost community call #0 (Recording)
The first call covered topics including updates on the Builder-specs for the Capella and Deneb (EIP-4844), and evaluating the readiness of relay and builder implementations. The conversation also focused on ensuring Capella hard fork readiness by coordinating with relay operators for a smooth transition during the hard fork process.

The group then moved to discuss the Optimistic Relay pull request, which enables bids to come in closer to the end of the slot. This is achieved by the relays not validating the blocks before forwarding them to the proposer, instead relying on collateral being put up by the builder.

A transcript of the meeting was published by @apriori.

MEV-Boost community call #1 (Recording)
The second call focused on the successful Sepolia hard fork, preparations for the upcoming Goerli hard fork, and necessary preparations for relay operators. Participants addressed block validation concerns and discussed communication channels for monitoring and sharing urgent issues.

The discussion from the previous call related to Optimistic Relay continued with @mikeneuder.eth sharing the roadmap for Optimistic Relay with the long-term goal to achieve enshrined PBS. A lively conversation ensued with concerns raised about latency wars, centralization, and collateral requirements.

Once again, @apriori shared a transcript of the meeting! :zap:

The third MEV-Boost community call is scheduled for March 30th, 16:00 UTC.

MEV-Boost development

MEV-Boost v1.5.0
A new stable version of MEV-Boost has been released with a focus on ensuring Capella readiness. It will also introduce getPayload retries which will mitigate the risk of proposers missing their slot solely due to transient network issues. v1.5.0 also includes dependency upgrades, and a transition to Go v1.20.

MEV-Boost-Relay v.0.16.0 -alpha1 (pre-release)
Similar to the MEV-Boost release, MEV-Boost-Relay v0.16.0 will feature Capella upgrades. Two extra timestamps collected in the block lifecycle will also be added.

MEV-Boost and Builder spec Capella upgrades
The Capella upgrade on mainnet is fast approaching and is now just weeks, not months, away. This follows the successful Sepolia upgrade on February 28 and the Goerli upgrade on March 14. Comprehensive documentation detailing all Capella-related changes has been published, covering all actors in the MEV-Boost stack, including block proposers, relays, and block builders.

The documentation for Running MEV-Boost-Relay at scale has also been revised with regards to Capella together with information related to pruning the builder submission table in the Postgres database.

MEV-Boost Relay data bulk downloads
The Flashbots MEV-Boost Relay’s bulk data downloads have been updated and automated, featuring over 400 million builder bids and more than 610k delivered payloads since the Merge until now.

Capella engine API

The newest Capella engine API includes a noteworthy addition: the local execution client now returns the payload value alongside the payload itself. This improvement will allow the consensus client to compare the locally built block with the block received from MEV-Boost.

As described in the Transparency Report from November, being able to compare the value of a locally built block with the value of a block from a third-party builder will open up new possibilities that give proposers greater flexibility in how they operate.

As described in The Cost Of Resilience, the validator could for example propose a locally built block if the difference between the block sourced from MEV-Boost and the locally built block is below a certain value, as opposed to deciding only based on the payload value from MEV-Boost with the min-bid parameter.

Research and discussions

Decentralized crypto needs you: to be a geographical decentralization maxi

In this post @phil emphasizes the dangers posed by latency games and underscores the necessity for geographical decentralization. The post examines trends in MEV that have the potential to centralize and weaken the core mission and value proposition of cryptocurrency. The post argues that the most important only exit from a future where power dynamics in cryptocurrency become centralized and predatory is through geographic decentralization. It then explores the relationship withbetween geographic decentralization and privacy, which is presented as the dominant economic phenomenon in the next decade of MEV evolution.

Backrunning Private Transactions Using Multi-Party Computation

In this article @ra explores the application of secure multiparty computation (MPC) to enable searchers to backrun users’ transactions while maintaining the confidentiality of both transactions and searcher strategies. By utilizing MPC multiple parties can collaboratively compute a public function while keeping their respective inputs private, revealing only the final result.


MPC-based Backrunning Private Transactions

This MPC setup shares similarities with SGX but differs in that the backrunning program is executed as a communication protocol between the user and the searcher, rather than as a program within an SGX enclave on the searcher’s machine.

If you wish to learn more, check out @ra presentation during the MEV Roast on Privacy and see the proof-of-concept with instructions for backrunning private transactions using MPC on Github.

Proposer Auction Simulations: Ideas Wanted!

@jolene is seeking ideas on alternative ways to design the market for block construction in MEV-Boost that maximize welfare. The collected suggestions will then be used in models and simulations to gain a deeper understanding of the current PBS market’s structure.

Check out the thread and contribute with your suggestions to help us improve the block construction market!

Multiplicity: A Gadget for Multiple Concurrent Block Proposers (aka Ending the Proposer Monopoly)

@eljhfx and Max Resnick expand on their post published on the Ethereum Research for the context of Flashbots. The post explores ways to enhance censorship resistance in blockchains and introduces Multiplicity as a practical gadget for allowing multiple concurrent block proposers.

The goal of Multiplicity is presented as a tool to build censorship-resistant on-chain auctions that will enable decentralizing block building and for MEV to be recaptured at the application level. It may also be used to achieve decentralized sequencing for rollups.

Cross-domain MEV with EigenLayer

In this post @Walt0x explores potential centralization and neutrality issues that could arise from cross-domain MEV and EigenLayer. It highlights risks like multiblock MEV, and explores mitigation strategies through SUAVE, application design and MEV-smoothing. The post is a shortened version of the original post published on the EigenLayer forum.

Suave Economic Security Models

In response to @Jon’s thread, @Hasu elaborated on the economic security models for SUAVE, noting that while rollups could be a long-term solution, SUAVE doesn’t substantially benefit from the safety and liveness guarantees a rollup would provide.

Hasu also emphasized the current focus on innovative transaction types and the programmable privacy framework instead of the specific architecture used for securing the chain.

Flashbots Research Proposals (FRPs)

Two FRPs have been completed since the previous report with published papers showcasing their results:

FRP-21: MEV in fixed gas price blockchains by @facuzeta

This research explores MEV opportunities within blockchains with fixed gas prices, focusing on the Terra Classic blockchain from September 2021 to the de-peg event in May 2022. The objectives included analyzing Terra Classic’s arbitrage characteristics, comprehending the strategies searchers use to optimize profits, examining timing-related aspects of arbitrage, and creating a dashboard to facilitate data analysis.

FRP-22: Quantifying the Impact of Frontrunning and Randomness on UX by @awmac
This FRP addresses the question “What happens when your order hits the mempool?”. The developed model offers a probabilistic perspective on mempool dynamics from a trader’s perspective on a Constant Function Market Maker (CFMM). The model integrates elements from game theory and queueing theory and is computable under certain conditions discussed in the paper.

A new FRP was approved earlier this month: FRP29: Research on MEV in L2 Blockchains by Erik Zhang and Cindy Jiang + FranklinDAO (Penn Blockchain) Research Committee, . The research will explore the current MEV landscape, with an emphasis on decentralizing sequencers used by L2s.

Flashbots is an open research organization and anyone can submit a research proposal and become eligible for a grant!

Flashbots at ETHDenver

Flashbots organized a series of MEV and research events at the Pirateship Hackerhouse and other venues during ETHDenver. Recordings of these events can be found below. See the ETHDenver forum post for additional details and discussion.

:date: In order to stay up to date on any upcoming events feel free to subscribe to the Flashbots Collective calendar!

Resources

Below is a collection of publications, podcasts, dashboards and other resources from the broader community posted since the last report on topics related to MEV, PBS and Flashbots. Check out the list from previous months for more fantastic resources and please feel free to share additional resources as replies below or in the bookmark-topics!

ETHDenver

Papers

Articles

Posts & threads

Podcasts & Videos

Dashboards

  • MEV Relay Dashboards and Alerts by Mevtrika provides real-time metrics and notifications to help monitor MEV-Boost Relays for Builders, Relays, and Node Operators.
  • Builder Landscape by Rated provides metrics of the builder market, including data on exclusive order flow.
  • The data behind mevboost.pics is now open source with info on every slot since the Merge, MEV-Boost and the mapped block Builders, Validators, and Relays.
  • The Ultra Sounds Relay dashboard have introduced a section for censorship with metrics on MEV-Boost relays and block proposers.

Other

  • The Auction Zoo by a16z is a Solidity implementation of a sealed-bid auction in an effort to “bridge the gap between auction theory and practice by showcasing Solidity auction implementations that demonstrate interesting theoretical properties or novel constructions.”
  • mev-bibliography by @mikeneuder.eth is a collection of links related to MEV-Boost, PBS, censorship resistance, metrics, MEV infrastructure and more.

Get involved :zap::robot:

At Flashbots, we research and build systems around MEV, and we would love to collaborate with you. We are a distributed organization with the principles of a pirate hacker collective, and we have several open positions. We also issue grants to external researchers doing work aligned with ours, please find out more in our Research repository.

Make sure to also look around on our forum and join the conversations!
Previous Transparency reports

15 Likes

Incredible work @Fred .
These are quickly becoming a must-read for anyone remotely interested in MEV.

4 Likes

Super helpful indeed.

2 Likes