A research collaboration among Sei Labs, Mysten Labs, and academic institutions has published Sedna, a user-facing protocol that replaces whole-transaction replication with verifiable, rateless coding for blockchain networks employing multiple concurrent proposer (MCP) consensus. The protocol directly addresses a practical trilemma among censorship resistance, low latency, and reasonable cost that has constrained transaction dissemination in MCP systems.
Modern blockchains increasingly adopt MCP consensus to remove single-leader bottlenecks and improve bandwidth utilization. However, MCP alone does not resolve how users should disseminate transactions to proposers.
Users face a binary choice: replicate full transactions to many proposers—sacrificing goodput and exposing payloads to MEV—or target few proposers and accept weak censorship and latency guarantees. Naive replication imposes significant goodput degradation; tolerating censorship from five proposers requires each user to send transactions to at least six validators, incurring an 80% worst-case goodput decrease.
Sedna’s Rateless Coding Architecture
Sedna replaces naive replication with a rateless erasure coding scheme. A sender first commits to the transaction payload and derives a transaction identifier. The protocol then shards the payload into verifiable rateless symbols and disseminates them across parallel proposer lanes.
The sender transmits small symbol bundles to a subset of proposers in each consensus slot. Execution follows a deterministic order once the chain has accumulated enough finalized symbols to decode the complete transaction. The protocol guarantees liveness and until-decode privacy, significantly reducing MEV exposure.
Bandwidth Efficiency and Information-Theoretic Bounds
The protocol approaches the information-theoretic lower bound for bandwidth overhead. Analytical results demonstrate a 2–3× efficiency improvement over naive replication. Sedna reduces per-lane bandwidth to a fraction of the payload size, whereas naive replication requires each proposer to receive the complete transaction.
The bandwidth savings derive directly from the rateless coding property: recipients can recover the original payload from any sufficiently large subset of symbols, not necessarily all symbols transmitted. This property eliminates the need for redundant retransmission across all proposers.
Until-Decode Privacy and MEV Protection
Sedna provides until-decode privacy: the payload remains private until the decode threshold is reached. No single proposer can view the full transaction details before finalization. This design reduces MEV exposure because front-runners cannot read pending transactions in time to act on them.
The cost of MEV protection under Sedna is 0.04% of the transaction value for realistic scenarios, at negligible latency cost. The protocol achieves approximately 90% of the privacy benefits of a full zero-knowledge layer while requiring substantially less implementation effort.
Subsequent research has examined incentive properties within Sedna. A cartel of lane proposers controlling a fraction of lanes could withhold addressed bundles, slowing symbol accumulation while privately pooling missing symbols. This creates a multi-slot information lead when the chain fails to accumulate the bundles needed for decoding by the honest horizon.
The protocol addresses this through PIVOT-K, a Sedna-native pivotal bundle bounty that concentrates rewards on the bundles that actually trigger decoding. An adaptive sender “ratchet” excludes lanes whose tickets were not redeemed, collapsing multi-slot withholding into a first-slot deficit. This mechanism reduces required bounty magnitudes by orders of magnitude.
Sedna requires no consensus modifications, enabling incremental deployment on existing MCP chains. The protocol operates at the user-facing layer, allowing senders to independently tune censorship tolerance per transaction.
Sei Labs has integrated Sedna into the Giga V2 whitepaper as a private transaction layer. The implementation targets sub-250ms finality while maintaining throughput of 200,000+ transactions per second. The upgrade is designed to deploy without regenesis or network downtime.
The sender performs a cryptographic commitment to the payload before encoding. Rateless symbols are verifiable, allowing recipients to validate symbol integrity without decoding the complete payload. The protocol defines κ = ⌈K/s⌉ bundles required for decoding, where K represents the total symbols needed and s represents symbols per bundle.
Full inclusion guarantees deterministic decode threshold achievement after t⋆ = ⌈κ/m⌉ slots, where m represents the number of lanes. The slack parameter Δ = t⋆m − κ ranges from 0 to m−1 and determines withholding vulnerability.
Practical Implications
The protocol resolves the user-facing trilemma by providing all three properties simultaneously. Trading applications require all three properties because latency effectively constitutes censorship—a brief delay can eliminate an opportunity even if the transaction eventually confirms.
Sedna’s efficiency improvements translate to direct cost reductions for users and systems. The protocol shifts the bandwidth cost from full replication to coded fragments, reducing the vulnerability to DoS attacks on goodput.
The research represents a formal treatment of coded dissemination in MCP systems, with proven liveness and privacy guarantees. The protocol’s information-theoretic efficiency and incentive compatibility mechanisms provide a foundation for deploying private transaction dissemination at scale.
FAQ
What problem does Sedna solve in blockchain transaction dissemination?
Sedna resolves the practical trilemma among censorship resistance, low latency, and reasonable cost that arises when users disseminate transactions to multiple concurrent proposers. Naive replication provides censorship resistance but sacrifices goodput and exposes payloads to MEV. Targeting few proposers preserves privacy but offers weak censorship guarantees. Sedna uses rateless coding to provide all three properties simultaneously
How does rateless coding improve bandwidth efficiency compared to naive replication?
Rateless coding allows recipients to recover the original payload from any sufficiently large subset of symbols, not necessarily all symbols transmitted. This eliminates redundant retransmission across all proposers. Sedna approaches the information-theoretic lower bound for bandwidth overhead, achieving a 2–3× efficiency improvement over naive replication.
What is “until-decode privacy” and how does it reduce MEV exposure?
Until-decode privacy means the transaction payload remains private until enough symbols have been finalized to decode the complete transaction. No single proposer can view the full transaction details before finalization. This prevents front-runners from reading pending transactions in time to act on them, significantly reducing MEV exposure.
Does Sedna require changes to the underlying consensus protocol?
What mechanisms prevent proposer cartels from withholding symbols?
Sedna implements PIVOT-K, a pivotal bundle bounty that concentrates rewards on the bundles that trigger decoding. An adaptive sender “ratchet” excludes lanes whose tickets were not redeemed, collapsing multi-slot withholding into a first-slot deficit. These mechanisms reduce required bounty magnitudes by orders of magnitude and establish incentive compatibility against partial and coalition deviations.
Stefan is the founder of Blockchaindose and a long-time tech enthusiast who first dove into crypto back in 2012, starting out by mining his first coins on graphics cards. Over the years, he shifted gears into active trading and investing. Today, his main focus is driving the platform forward and showing people how crypto can actually be used in the real world—beyond just the speculation.
That blend of early-adopter roots, hands-on market experience, and a practical vision makes his perspective a solid anchor for the team.
