The MEV conversation until this point has been focused almost exclusively on Ethereum, however this series explores MEV within the Solana and Cosmos ecosystems. This article will highlight a diverse range of approaches to address MEV and associated incentive alignment challenges, including comparative analysis of tradeoffs. By broadening the conversation, we can learn from the various experiments being run in the wild to better anticipate potential problems and ultimately design better solutions.
There is no shortage of threads making the case for why MEV is an important topic (from centralization to network performance and beyond), but the way MEV reifies the intrinsically social nature of blockchains may still be under-appreciated. The technical solutions that we design and deploy have and will continue to be guided by fundamentally political questions: Who are the stakeholders? Who should be the stakeholders? What degree of centralization or censorship risk is acceptable? Should we redistribute wealth? If so, how? It is well known that governance is often messy, often characterized by intense battles between competing interests, and taking on forms that reflect the unique qualities of the societies that give rise to them. In that sense, Part I of this series can be seen as a political analysis of MEV as much as an overview of emerging blockspace market structures.
On Ethereum, the rise of financial applications became the catalyst for more widespread research and understanding around MEV. The now seminal Flashboys 2.0 paper published in 2019 not only coined the phrase “MEV,” but also shed light on how exactly this phenomenon was manifesting itself onchain. It was the first time that most of us were exposed to concepts like priority gas auctions (PGAs) or searchers. Shortly thereafter, Flashbots was founded to tackle the many incentive alignment challenges associated with MEV head-on. They originally put forth three core goals to work towards:
Illuminate the Dark Forest
Democratize Extraction
Distribute Benefits
Having since achieved over 90% network adoption on both of their major product releases, it’s fair to say that these goals also evolved into the guiding principles that have largely shaped the development of Ethereum’s blockspace market.
Flashbots Auction (mev-geth + mev-relay) created a private transaction pool & off-chain sealed bid auctions. This allowed searchers to express more granular preferences for tx ordering in the form of bundles, as well as bid for inclusion at the top of the block without clogging Ethereum’s public mempool with failed arbitrage attempts and driving gas fees up severely.
Flashbots Protect is an RPC endpoint that can be added to consumer wallets like MetaMask. This extends to benefits of bundles to regular users.
MEV-Boost was designed in anticipation of The Merge, as MEV was expected to become an increasingly centralizing force in Proof-of-Stake. It was the first implementation of proposer-builder separation (PBS), which aims to separate the role of building a block from the role of proposing a block by creating a more decentralized, competitive market for block-building.
MEV-Share (unreleased) outlines designs for a protocol that matches transactions from users, wallets and/or applications with searchers in a private & permissionless fashion. Decentralizing access to order flow and introducing the notion of programmable privacy are among the key long-term goals here, however the matchmaker function will begin as a trusted role operated by Flashbots.
SUAVE (unreleased) is an upcoming release which aims to address the remaining centralizing vectors of (1) exclusive orderflow and (2) cross-domain MEV. Minimal detail has been shared about the nature of the project, however it appears that SUAVE will serve as a decentralized mempool & sequencing layer for the EVM ecosystem. It will feature an encrypted mempool, as well as introduce a new actor to the MEV supply chain: executors – who will compete to provide users with the best execution.
Much like Ethereum, the growth of Solana’s financial sector attracted MEV activity that started to degrade the end-user experience. However unlike Ethereum, where gas fees became prohibitively expensive, Solana’s problems were caused by a combination of (1) extremely low gas fees, (2) the absence of a fee market and (3) a sub-optimal transaction propagation protocol that didn’t offer much optionality around discouraging network spam. In moments of intense market volatility or popular onchain events like NFT mints, the amount of spam transactions being sent to the network was sometimes so significant as to cause network outages.
Some might wonder why the Flashbots product suite couldn’t simply be repurposed for Solana, however it's important to consider that Solana’s architectural design is distinct to Ethereum’s in ways that necessitate a unique approach. Key differences include Solana’s speed (400ms between blocks), unique data propagation protocol, mempool-less transaction forwarding, localized fee markets and parallelized transaction processing, just to name a few.
In 2022, Jito Labs entered the picture to take an ecosystem-specific approach to providing Solana with much-needed MEV infrastructure. Like Flashbots, they set out with a few key goals in mind:
Minimize the negative externalities of MEV
Prevent centralization
Distribute the rewards of MEV
Since then, Jito has been the dominant MEV solutions provider on Solana, and has released a number of products including:
Jito-Solana was the first third-party Solana client, optimized for efficient MEV extraction. Much like mev-geth, this allowed validators to support transaction bundles and was built to work seamlessly with the Jito Relayer and Jito Block Engine.
Relayer is meant to provide validators with a layer of protection between their TPU (transaction processing unit) and network spam. Validators can run their own relayer or use a version hosted by Jito Labs.
Block Engine is essentially a high performance block builder – it runs a sealed bid auction for blockspace, and forwards the most profitable bundles to the current leader for immediate execution. The block engine is also globally distributed to provide open access to low latency.
Searcher Tools:
Jito Mempool -- Despite Solana not having a mempool in the traditional sense, this allows searchers to subscribe to “accounts of interest” and atomically extract MEV from their transactions using bundles, allowing for a more proactive approach to searching.
ShredStream -- Run on Jito-Solana clients, this sends shreds directly to the locally connected block engine, allows searchers to access shreds forwarded by leaders and reduces the latency by hundreds of milliseconds. (Shreds are fractions of a block (the smallest unit) which are constantly being emitted as validators produce blocks. They are distributed to the network based on a stake-weighted shuffle, so that servers with a higher stake are likely to receive shreds more quickly, which can bring meaningful advantages when engaging in high frequency trading.)
MEV Payments & Distribution enables validators to seamlessly distribute MEV rewards to their stakes in the form of airdrops.
Chorus One, one of the largest node operators across all major crypto ecosystems, also recently released a whitepaper outlining a prototype for Solana-MEV – a modified client designed to further decentralize extraction for validators without adding unwanted latency to the system. The client would make it easy for validators to check for potential MEV opportunities after each batch of user transactions and insert their own to capture the value.
Despite having arguably the most nascent DeFi ecosystem, Cosmos promises to be fertile ground for experimentation in both blockspace market design and cross-domain MEV. Unlike Ethereum, where trading & lending volumes are already in the billions of dollars, and Solana, where low-latency financial applications were the primary imaged use case, Cosmos seems to have been slower to the MEV conversation. There are many possible reasons for this, but the most obvious explanations are Tendermint clients using first in first out (FIFO) ordering by default as well as a lack of financial activity.
It was not until the release of Osmosis in late 2021 and even more, the collapse of Terra in May 2022, that anyone began to notice and measure meaningful amounts of MEV being captured within the Cosmos ecosystem. Combine these events with bear market conditions like a collapse in fees, and it becomes easy to see why validators began to consider other options in an attempt to remain profitable. Foreshadowing this, a number of Cosmos-native MEV solutions providers have entered the scene, most notably Skip Protocol and Mekatek.
Zenith is meant to create an open market for block building within the Cosmos. Searchers can submit transaction bundles and compete for priority inclusion within blocks, while validators can outsource block-building to Zenith and sell their blockspace for maximum profit.
Mev-Tendermint is a modified version of Tendermint that allows validators to accept transaction bundles and introduces a seal bid auction for inclusion at the top of the block (Skip does not build the entire block, only the top).
Skip-Select brings Cosmos-style sovereignty to MEV by allowing for fully configurable, governance-driven blockspace auctions. It makes it easy for validators to decide how to split MEV rewards, what % of building for a given block should be outsourced to Skip, whether or not to protect blocks from frontrunning/sandwiching, and more. Skip-Select is also laying the groundwork for a future where in-protocol MEV preferences are voted on and implemented via onchain governance, which is uniquely enabled by the Cosmos SDK and ABCI++ (application-blockchain interface).
Skip Secure is very similar to Flashbots Protect – private transaction RPC that end-users and frontends can leverage for private execution.
Proto-Rev – arguably Skip’s most exciting product – refers to custom-built modules for enshrining certain MEV preferences into the core protocol. The first implementation of proto-rev was proposed to help Osmosis capture some meaningful amount of arbitrage-based MEV rewards internally, but the service will be available on a chain-by-chain basis.
Although there is less publicly available information on the project, FairBlock is building IBE-based solutions (identity based encryption) to address MEV-related challenges, such as ensuring end-user access to pre-execution privacy and best execution. The team plans to leverage Interchain Security from the Cosmos Hub for their consumer chain, which will be used to manage and distribute the validator decryption keys.
The success of PBS – The notion of proposer-builder separation has spread beyond Ethereum, as some version of it now exists within both Solana and Cosmos.
Latency Wars – In part because of Solana’s network architecture, Jito Labs takes a very latency-sensitive approach to MEV extraction, which could be especially compelling for people running servers and/or validators outside USA/Europe (or anywhere connectivity to the rest of the cluster is less than ideal). Latency could also be an interesting vector for aspiring Flashbots competitors to experiment with, although such an approach would likely come with less than ideal implications for decentralization and incentive alignment.
Enshrined Solutions – Due to the level of sovereignty and autonomy that Cosmos communities have over their tech stack, it is generally much easier to enshrine MEV solutions (such as PBS) into the core protocol. There are a number of reasons for this, one being the political dynamic of achieving social consensus among single-issue voters (appchains) vs. factions of competing interests (general L1s). As challenging as Cosmos governance can often be, the fact that it can take place onchain and the fact that there is no need to consider how changes will affect other applications can be a meaningful accelerant to the implementation process. While Flashbots and many Ethereum core developers have expressed a desire to enshrine PBS into the Ethereum protocol, such an upgrade will no doubt require a significant amount of technical work and social consensus building. On the technical side, innovations such as the latest version of ABCI++ allow for new possibilities with respect to how Cosmos-based applications communicate directly with the consensus layer, allowing for features like threshold encryption, for example. For this reason, we should expect most of the innovation around in-protocol solutions to come out of the Cosmos.
Cross-domain MEV – As noted by Flashbots in their SUAVE primer, incentives to capture cross-domain MEV are expected to ramp up in the coming years, and could pose serious threats to the economic security models of various ecosystems. Because Cosmos was built for cross-chain on day one (over 50 chains connected by IBC) and because knowledge of how to run one version of a Tendermint client likely lends itself well to learning how to run others, it may be uniquely exposed to the centralizing forces of cross-domain MEV. There are already large entities that run validators on multiple Cosmos chains, and if left unchecked, it’s easy to imagine this leading to a world where only a handful of and extremely well-capitalized validators control meaningful stake across various chains.
Part II will explore the implications of the above for end-users and applications, specifically: Focus shifting from blockspace → orderflow
The design space for mev-aware applications & wallets
Incentives for private orderflow and vertical integration
MEV & Privacy
Huge thank you to Barry from Skip, Drew from Brevan Howard and Hasu for reviewing this piece & providing thoughtful edits.
If you’re thinking about or building in any of these verticals, want to chat or collaborate on further writing, feel free to reach out via DM or to natalie@volt.capital.
Over 100 subscribers
natalie
