SOLANA·HUB
DE
Infrastructure

Constellation: Multi-Proposer Proposal for Solana Block Construction

Helius Research has published Constellation as a protocol proposal for Multiple Concurrent Proposers (MCP) on Solana. Anza has published a complementary roadmap position. Overview of the mechanism and open problems.

SOLANA·HUB Editorial · ·
Source: Helius Research

Overview

Helius Research has published Constellation, the first formal protocol proposal for Multiple Concurrent Proposers (MCP) on a production blockchain. Anza has separately published a roadmap post on the adoption of similar mechanics.

What Constellation Proposes

From the Helius Research post:

  • Instead of a single block leader deciding alone on transaction order, approximately 16 proposers work in parallel on a 50-millisecond cycle
  • Proposers assemble transactions into erasure-coded “pslices” and distribute them to approximately 256 attesters
  • Attesters cryptographically sign which transaction sets they observed
  • If enough attesters witness a pslice, the leader can no longer selectively exclude transactions without producing a block the network will reject

The central property: selective censorship resistance. Either all fee-competitive transactions in a cycle are included, or none are.

Open Problems (per Helius itself)

Helius names two unresolved points in the research post:

  • Content-visible ordering remains unsolved. Transactions are visible to all proposers on receipt — the multi-proposer architecture may actually widen this attack surface
  • Time-based latency games are described as “unpunishable” under the current design

Anza’s Parallel Roadmap Post

Anza published a separate post on March 25, 2026 (Anza Blog: Solana Constellation — Fair Internet Capital Markets) addressing the multi-proposer proposal from Anza’s perspective. Anza CEO Brennan Watt mentions MCP additionally in the Anza26 roadmap post: “In 2026, we’ll ship an initial MCP version focused on enforcing transaction ordering within a batch in-protocol.”

The exact relationship between the Helius Constellation spec and Anza’s MCP implementation is subject to ongoing discussion within the Solana Foundation.

How Constellation Relates to Alpenglow

Constellation and Alpenglow address different Solana layers:

  • Alpenglow = consensus layer (how validators agree on finalized blocks)
  • Constellation = block construction layer (who initially assembles blocks)

They can technically coexist. Alpenglow finalizes blocks regardless of whether they come from a single leader or multiple concurrent proposers.

Status

  • Constellation is a research-stage proposal
  • A concrete mainnet implementation timeline has not been publicly named by either Helius or Anza
  • The Solana Foundation and validator communities are continuously discussing refinements

What to Watch

  • How the Helius spec and the Anza implementation roadmap converge
  • How validators operationally integrate the additional “proposer/attester” roles
  • How the unresolved problems (content-visible ordering, latency games) are addressed in subsequent spec versions
  • When a formal SIMD emerges from Constellation

Source

#constellation #mcp #anza #helius #block-production