Two central design goals for blockchain protocols are maximizing censorship resistance and minimizing latency. These goals are in tension: Censorship resistance requires additional rounds of communication, while low latency requires minimizing communication rounds. This tradeoff is especially acute if blockchains are to support efficient onchain markets: Market participants need both low latency and robust protection against proposer censorship.
The resolution is that censorship resistance does add protocol rounds, but it reduces the latency users experience when they require censorship-resistant inclusion.
One line of work focuses on adding strong censorship-resistance guarantees to blockchain protocols. For example, Ethereum is considering FOCIL / EIP-7805, and Solana is considering Constellation (also see MCP). Both are based on a simple two-round inclusion-list idea: before a block is proposed, validators collect transactions that the next block is not allowed to ignore. Similar proposals are being considered across the blockchain ecosystem.
Another line of work focuses on reducing latency to the minimum possible. Formally, this is measured by good-case latency: how quickly a blockchain can commit a block when the proposer is honest and the network is well-behaved. Our recent analysis shows that, in standard Byzantine Fault Tolerant (BFT) consensus, 3 rounds is the minimum when more than one-fifth of validators may be faulty. Most production protocols are designed for a stronger fault model, usually up to n/3 faulty validators.
But standard BFT hands the proposer unchecked power over block content. A malicious or economically motivated proposer can simply exclude a user’s transaction, a form of censorship that, in blockchain systems, is the root of much MEV-related harm. In a previous post, we described Strong Chain Quality (SCQ) and the inclusion-list mechanism that makes censorship resistance possible: The next block must include transactions submitted by any honest validator.
In this post, we characterize the fundamental cost of that guarantee: Any censorship-resistant BFT protocol requires at least 2 extra rounds beyond the standard BFT baseline, for a minimum good-case latency of 5 rounds. The key point is that these extra rounds buy a stronger user-facing guarantee: Censored transactions no longer wait through an arbitrary number of blocks.
This isn’t an artifact of any particular protocol design. It is a mathematical lower bound.
Standard BFT: The proposer’s monopoly
In a partial synchrony BFT protocol, a designated proposer both constructs the block and drives consensus. The system’s key performance guarantee is:
Good-case latency: the number of communication rounds until all honest validators decide when the proposer is honest and the network is synchronous.
When the number of faults exceeds n/5, the best achievable good-case latency is 3 rounds; obtaining two rounds is provably impossible. Most modern BFT protocols (e.g., PBFT, Tendermint, Casper, Simplex) indeed have a 3-round good-case latency.
The proposer’s monopoly over both construction and proposal of the block creates a censorship problem: A malicious proposer can ignore inputs from other parties. In blockchain systems, this corresponds to validators excluding user transactions for profit.
Decoupling construction from proposal
The natural fix — which we described in the SCQ post — is to break the proposer’s monopoly by separating two roles:
- An input party I holds the value to be committed and broadcasts it to all validators before the block is proposed.
- An expediter E drives consensus, but cannot censor I‘s transaction, because all validators have already received it and committed to including it via their inclusion lists.
This separation defines a censorship-resistant protocol, with the following formal guarantee:
Censorship resistance: If the input holder is honest and the network is synchronous, its input value is included in the next block.
The performance question becomes: What is the best achievable good-case latency when the expediter, E, is honest? At first, 4 rounds seems plausible: The input holder, I, sends its value to E in round 1, then E runs a standard 3-round good-case BFT protocol. But there is a catch: A malicious E could simply claim it never received I‘s value, and essentially censor I. The fix is for I to broadcast its value to all validators, who then add it to their inclusion lists. This consumes an additional round before E can begin driving consensus, suggesting that 5 rounds is the right answer.
Our new result confirms this:
Theorem: Any censorship-resistant BFT protocol in partial synchrony requires good-case latency of at least 5 rounds when the number of faults exceeds n/5.
Why 4 rounds is impossible: The proof idea
The intuition is that censorship resistance requires validators to hear from the input holder before they can safely accept a block that omits its transaction. A 4-round protocol would leave only two rounds for the honest expediter to drive agreement after this censorship-resistance step, but the 3-round lower bound for Byzantine broadcast shows that two rounds are not enough in this fault model. Therefore, the tempting 4-round design is impossible; the best achievable good-case latency is 5 rounds.
The full proof, which formalizes this intuition using the AS-FFD method and a reduction to the 3-round lower bound for Byzantine broadcast, is in the paper.
What this means for protocol designers
At first sight this finding may seem daunting: Censorship resistance costs two rounds relative to standard BFT, taking good-case latency from 3 rounds to 5 rounds.
But without censorship resistance, low latency doesn’t matter much. A latency-sensitive transaction that is censored for even one block waits through the censored block and then waits again for the block that finally includes it. If it is censored for several blocks, the delay grows with every missed slot. In a censorship-resistant system, that delay is capped by the 5-round inclusion guarantee in the good case.
In this sense, censorship resistance does not make the fastest uncensored path faster. It makes inclusion predictable: It reduces effective censorship-resistant latency from “however long censorship lasts” to 5 rounds.
***
The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.
You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investment-list/.
The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures/ for additional important information.