Anza proposes Constellation to end Solana's single-leader monopoly with concurrent proposers on 50ms cycles

5 recorded changes
Want your article here?
Promote with Leviathan News
5 recorded changes
Want your article here?
Promote with Leviathan NewsAnza, a core Solana client development team, has published Constellation, a protocol design that replaces Solana’s current single-leader block proposer model with multiple concurrent proposers (MCP) operating on 50ms economic cycles. Under Solana’s existing architecture, each ~400ms slot is controlled by a single leader who decides which transactions are included and in what order, creating a temporary monopoly over transaction ordering and associated MEV (maximal extractable value). Constellation reframes this as a market-structure problem and attempts to dismantle that monopoly by splitting transaction proposal and ordering across a broader set of participants. In Constellation, 16 proposers per slot collect user transactions and, every 50ms, convert them into erasure-coded fragments (pshreds/pslices) that are sent to 256 attesters, which timestamp and attest to what they have received. If a proposer’s data reaches a quorum of attesters, the current slot leader is then cryptographically obligated to include those transactions in the next execution batch, sharply constraining their discretion over censorship and ordering while still leaving them responsible for assembling the final block. This structure aims to make censorship resistance and fairer ordering protocol-enforced rather than policy-based, while preserving Solana’s performance and integrating with the existing Alpenglow execution pipeline. Economically, Constellation introduces a 50ms protocol-enforced tick rate that Anza argues is the fastest censorship-resistant economic cycle among production blockchains, and it modifies fee flows without radically altering validator incentives. Validators still earn fees for block production, but transaction revenue is split into inclusion fees paid to proposers who include transactions and priority fees distributed by stake weight across validators, reducing the concentration of value in the temporary slot leader. The proposal has sparked active discussion in the Solana community: supporters highlight improvements to fairness, MEV structure, and market quality for activities like perpetuals trading, while critics warn of possible short-term revenue impacts and complexity for validators. Constellation is currently a design proposal open for community feedback and is expected to progress via a formal Solana Improvement (SIMD) process before any mainnet adoption. "entities":["Anza","Solana","Constellation","Multiple Concurrent Proposers (MCP)","Solana validators","Solana block leader","Alpenglow","MEV (Maximal Extractable Value)","pshreds/pslices","attesters","Solana Foundation","Ilan Gitter","Helius","Gate Learn / Gate.io","Chainflow"]}`
AI-generated background, compiled from web sources — not editorial content.
Loading related coverage…
Loading comments…
Danicjade
This will be a massive win for the Solana ecosystem
Danicjade
This will be a massive win for the Solana ecosystem
Danicjade
Who tf is Anza. Never heard of them before
Benthic
16 concurrent proposers feeding erasure-coded slices to 256 attesters every 50ms is an insane coordination requirement — clock sync alone at that granularity will punish anyone outside a few major data center regions, which creates a tension with the geo-distribution goal. Bigger question is what this does to Jito's ~$1M+/week tip economy: if the leader loses ordering discretion and the protocol enforces deterministic sequencing, sandwich bots don't just get harder — the entire MEV supply chain that validators have been monetizing gets protocol-killed, and you have to wonder how many validators stay economically motivated without that revenue stream subsidizing their hardware costs.

𝕏/@solana ·

𝕏/@SolanaFloor ·

𝕏/@metaplex ·

𝕏/@BackpackOnchain ·

𝕏/@platis_e ·

𝕏/@domaprotocol ·

𝕏/@solana ·

𝕏/@SolanaFloor ·

𝕏/@metaplex ·

𝕏/@BackpackOnchain ·

𝕏/@platis_e ·

𝕏/@domaprotocol ·