Flashbots Transparency Report: May & June 2023

Welcome back to another edition of the Flashbots Transparency Report. Before we jump in I want to mention a slight change to the structure of these reports going forward – In order to streamline the content, and make the reports more digestible, I’ll be breaking out the ecosystem highlights into a separate MEV newsletter. This means that the resource list, typically included at the end of these reports will be published in this separate, weekly post to ensure the information remains timely and relevant. Keep an eye on the forum for the first edition to be published next week!

The Transparency reports will continue to be published and contain the latest development, research, discussions, and events related to Flashbots. This edition will cover highlights from May and June. Enjoy!

Redistributing MEV through MEV-Share

We’ve now surpassed 2 million blocks since the Merge, with over 20 million ETH staked on the Beacon chain. Proper-Builder separation through MEV-Boost has enabled a total of 220k ETH to be distributed to proposers, helping the validator set on Ethereum stay decentralized and resilient to capture by sophisticated actors.


From Flashbots Transparency Dashboard: Cumulative ETH paid to proposers since the Merge.

As we continue our mission to democratize and distribute MEV, we look to build protocols that redirect much of this value to back users by granting them more control and bargaining power. Just prior to the last report we announced MEV-Share Beta, and at the end of last month, a record-breaking refund of 10.8 ETH was refunded via a single blind backrun.

If you’re interested in learning how Searching on MEV-Share works in more detail, check out the walk-through of simple-blind-arbitrage hosted by @scott, @brock & @bert. The session delved into how simple-blind-arbitrage can be used to blindly submit atomic arbitrage opportunities through MEV-Share in order to equip searchers with the context needed to build out their own strategies leveraging programmable privacy.

The matchmaker-ts library used by searchers to send bundles to the Matchmaker has received several updates with new features since the last report. As a searcher you’re now able to use the tx_hash hint to communicate with other searchers what transaction you’re bundling (and get a share of their profits if they backrun you!).

Another notable new feature is the deployment of our SSE history API which allows users to look up all historical events shared by the Matchmaker. This provides a quick way to get a feel for the data on MEV-Share, and offers plenty of events to backtest strategies with. For more information on these new features, see the tweet thread by @brock and check out the examples in the matchmaker-ts repository.

Note that the MEV-Share protocol is under active development with new features being added on a regular basis. We welcome and encourage all feedback and questions and invite relevant actors to join the Decentralized Orderflow Working Group to be part of the research and development of the decentralized future of the MEV supply chain.

See our documentation for information on how to connect as a user, and how to integrate MEV-Share into your wallet or dapp. If you’re a searcher looking to get started with MEV-Share and have any questions or feedback, feel free to reach out here on the forum or through Discord!

MEV-Boost

As we approach the one-year mark since the Merge, a period of stabilization can be observed in both the relay and builder markets. Currently, four relays account for 80% of the total market share, while five builders hold an equivalent 80% market share.



From mevboost.pics: Both the relay market (top chart) and builder market (bottom chart) have stabilized in recent months.

The newly published paper Time to Bribe: Measuring Block Construction Market by Anton Wahrstätter, Liyi Zhou, Kaihua Qin, Davor Svetinovic, Arthur Gervais provides a deeper and more comprehensive analysis of these markets and the financial dynamics within the MEV-Boost ecosystem. The paper is packed with interesting data and takeaways, if you’re interested in these topics I recommend giving the paper a read!



From Time to Bribe: Measuring Block Construction Market.

The Beacon chain losing finality

On May 11th and 12th Ethereum experienced two separate events where finality was unable to be reached for 3, and 8 epochs respectively. The root cause of the problem was a large number of missed blocks, which then resulted in nodes attesting to old target checkpoints. Compounding the problem, nodes were subjected to high stress due to the processing of a large volume of deposits. Some nodes were overwhelmed by this surge, which ultimately resulted in the loss of finality.

Consensys Teku and Offchain Labs Prysm released fixes that eased the high resource usage nodes would experience during these scenarios. Offchain Labs also published a post-mortem analysis outlining the incident, what went wrong, where we got lucky, the fixes, and the lessons learned.

The temporary loss of finality had minimal impact on end-user transactions and the chain recovered without any intervention in both incidents. The second event caused inactivity leak penalties to activate for the first time on Beacon chain mainnet. Additional information on inactivity leaks can be found in the Upgrading Ethereum handbook by Ben Edgington. Information on the impact the event had on the p2p network and on users’ on-chain experience can be found in the detailed write-up by soispoke: Cascading Network Effects on Ethereum’s Finality.

Preparing for Cancun-Deneb

The upcoming Cancun upgrade will include EIP-4844: Shard Blob Transactions which introduces a new transaction format and blobs. An initial 4844 update for mev-boost builder-specs has been merged and @terence published Ethereum validator & builder interaction for Deneb that summarizes the updates coming in Deneb that require new validator and builder interactions.

In addition to the preparations for Cancun, a new version of mev-boost, v1.6, has been released. This version is optional but recommended and includes additional data availability, logging improvements, and other minor updates.

mev-boost-relay

Our open-source implementation of the MEV-Boost relay has received a couple of updates since the last report. Below are some of the more noteworthy changes, see the linked release notes for a more comprehensive list of changes:

Besides the changes listed above, there’s been a number of significant performance improvements made. @metachris published an overview in two parts detailing these changes: Part 1 & Part 2. Thanks to these recent improvements, zero-downtime deployments are now possible for MEV-Boost relays.

MEV-Boost-Relay Database Migration Guide by @metachris provides steps for managing database growth and reducing operating costs for the MEV-Boost-relays. The post outlines a custom strategy to migrate data to a new, smaller database. It also includes scripts and queries for those who wish to implement a similar migration. For more handy tips related to running a MEV-Boost relay, see Running MEV-Boost-Relay at scale.

Disclosure: boost relay vulnerabilities reported by user @nrv by @metachris describes the details, resolution, and impact of the three vulnerabilities in the mev-boost-relay codebase, found by nrv in April. The issues were resolved promptly and no exploitation was found. nrv asked for the bounty reward to be donated to Médecins Sans Frontières.

Unbundling attacks on MEV relays using RPC by Michael Sproul from Sigma Prime discloses another unbundling attack on MEV-Boost relays that exploits the handling of equivocating blocks received via RPC. A temporary fix to mitigate the vulnerability was included in Lighthouse v4.1.0 with further fixes being considered. The client diversity of Ethereum would have made this attack hard in practice as clients deal with this issue differently.

Prio-load-balancer

The prio-load-balancer which allows relays to manage permissionless builder submissions via high and low-priority queues, has undergone several updates since the last report. These updates have significantly improved logging, providing more relevant information about request performance, including detailed timings and the general state of the system. For more details, see the release notes linked below:

MEV-Boost community call #4

(Recording)

The latest MEV-Boost community call discussed a range of topics, including the ongoing work on EIP-4844 implementation and the next steps for optimistic relaying, with an increase of the collateral from 1 to 10 ETH. The discussion also explored potential funding structures for relays, such as a “Relay Guild”, with a roundtable discussion planned for the upcoming EthCC in Paris.

Keep an eye on this forum, and the Flashbots Collective calendar to get notified of upcoming calls!

Research and discussions

Quantifying Info I: Quantifying The Information Leakage of Computing on Encrypted Transactions by @quintus presents a problem in performing computation on private information where an adversary tries to identify an unknown function from a set of functions using limited queries. The post concludes by asking for literature on this subject, measures of adversary’s learning, and the best policy under time constraints.

[WIP] Quantifying Info II: Quantifying The Impact of Economic Impact of Information by @quintus is a WIP follow-up that introduces economics into the problem by considering the value or impact of the leaked information. It uses two examples to illustrate that how much an adversary benefits from leaked information can depend on the specifics of that information. The post then changes the original problem by including rewards for both the adversary and the one sending the information. This change enables actors to measure the consequences of information leakage and suggests that some leakage might sometimes be useful.

Intent-Based Architectures and Their Risks by @quintus and Georgios Konstantopoulos discusses the concept of “intents”, allowing users to express a desired outcome while outsourcing the task of best achieving that outcome to sophisticated third parties. The post highlights potential risks associated with intent-based designs, including issues of centralization, trust, and opacity. The article further explores the concept of “intentpools,” which are mechanisms for intent discovery and matching and discusses the challenges in designing such systems.

[Re]distribution study circle spearheaded by @guayabyte to study different redistribution approaches has started with two sessions being hosted so far.

  • The first session included an introduction to MEV and a discussion how to map MEV to power asymmetries in the dominant economy.
  • The second study circle was hosted by @QZ and had a focus on deflationary economies and burning money as a mechanism to redistribute wealth.

A discussion has been initiated on the forum to collect suggestions on topics for upcoming meetings.

Automated Discovery of Potential MEV Opportunities with Contract Analysis by @elegant.song6200 introduce work done to automatically identify potential front-running and MEV opportunities in smart contracts before they are interacted with. The work intends to provide early warnings to users about potential MEV exposure, and simplify the decision process regarding private transaction submission.

Visualization for builders reputation by @vita discusses the risk of block builder centralization and censorship resistance as a result of private order flow. A potential solution is presented by making block builder reputations visible, making it easier for searchers to send to multiple builders. The solution is compared to the behavior of searching for reviews for hotels or restaurants to reduce the likelihood of encountering low-quality service.

Towards Trustless Optimistic Relaying by @SFYLL discusses ways to reduce trust towards MEV-Boost relays. The post suggests leveraging zk-proofs to minimize trust assumptions towards Optimistic relays. The post includes a link to a Github repo with code and more details on the problems and upsides of the proposed approach.

Flashbots Research Proposals (FRPs)

FRP-24: Quantifying MEV on L2s

@Eru_Iluvatar and @0xpanoramix published results from their research with data on atomic token arbitrages on Optimism and Arbitrum, with results from Polygon to be released at a later stage.


From FRP-24: Quantifying MEV on L2s: MEV Profit by date for Optimism (first chart) and Arbitrum (second chart).

@pillarbear responded in the thread introducing a Dune Dashboard they made with a less exhaustive methodology but with similar results.

FRP-26: Credible Commitments via Open Games

FRP-26 has been completed by 20squares with the team presenting a model built in their open games engine that implements:

  • The mental experiments that can be found in Xin’s slides about credible commitments, and
  • A more involved example involving a basic AMM.

FRP-27: Auction Simulations under PBS

The other FRP by 20squares on simulating various auction mechanisms for PBS has also been completed. The model allows to simulate and check equilibrium strategies for different kinds of auctions, and it is designed to be modular and extendable.

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

Events, talks & podcasts

Bell Curve Season 4

The fourth season of the Bell Curve podcast dedicated to the topic of MEV with @hasu as co-host has been wrapped up. Since the last Transparency Report, 3 new episodes on MEV have been published, bringing the total number of episodes to 9:

  1. MEV In The Cosmos | Henry de Valence, Barry Plunkett
    Barry Plunkett and Henry de Valence discuss MEV in Cosmos compared to Ethereum, the benefits of AppChain sovereignty, the importance of privacy, and more.

  2. MEV 2.0 - Order Flow Auctions & Privacy | Quintus Kilbourn, Barnabé Monnot
    @quintus and Barnabé Monnot join the podcast to discuss OFAs, ePBS and the future of MEV. The conversation then goes into the risks of vertical integration and how we can build systems to mitigate centralization points within the MEV supply chain.

  3. An MEV Masterclass
    In the final episode of the season Mike and Hasu discuss their takeaways from the season, what questions remain unanswered and where MEV markets will go next.

MEV-Week in Paris

Flashbots will host a series of discussions and events during EthCC in Paris this year. Details of these events are still being finalized, keep an eye out on the forum as we’ll post more information soon.
If you’re in Paris and would like to chat – don’t hesitate to reach out!

: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!

Papers

Articles

Posts & threads

Podcasts & Videos

Other

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

9 Likes