Optimum has launched its mump2p data-propagation protocol on Ethereum mainnet, moving the technology from testing into live validator infrastructure. The company says 16 node operators are already using mump2p across validators representing at least $3.4 billion in staked ETH, including Blockdaemon, Everstake and ether.fi validators.
According to Optimum’s October 20 mainnet announcement, mump2p delivered beacon blocks 3.3 times faster than Ethereum’s standard propagation path across a sample of 50,292 slots. The company measured a 335-millisecond average advantage using data from its connected nodes and Xatu sentry observations of conventional libp2p propagation.
Those results remain Optimum’s own measurements rather than an independent network-wide benchmark. The company reports that mump2p accelerates roughly 85% to 90% of observed blocks by between 2.5x and 3.5x, while also reducing variation in arrival times, particularly for nodes at the slower edge of Ethereum’s distributed network.
mump2p Changes Block Routing, Not Ethereum Consensus
mump2p operates at Ethereum’s networking layer rather than changing its consensus rules. According to Optimum’s technical documentation, the protocol uses Random Linear Network Coding to split block data into coded shards that peers can forward before receiving a complete copy, allowing downstream nodes to reconstruct the original message from sufficient independent pieces.
That architecture is designed to reduce propagation latency across geographically distributed validator infrastructure. Faster block arrival can give validators more time to identify the correct chain head and submit timely attestations, potentially improving consensus-layer reward capture without changing Ethereum’s underlying proof-of-stake mechanism.
Optimum reported one operator-specific example in which wrong-head votes declined 51% during the week after mump2p deployment compared with the preceding week. That 51% result is a single case study rather than a fleet-wide performance figure, so broader validator adoption will be needed to determine whether similar attestation improvements persist across different operators and network conditions.
The deployment arrives while Ethereum is putting increasing emphasis on networking performance as part of its broader Layer 1 scaling strategy. Glamsterdam development has already exposed peer-to-peer latency and recovery issues during testing, while future capacity increases depend partly on blocks and consensus data reaching validators reliably within strict timing windows.
Signal Opens Broad Access While Accelerate Targets Larger Operators
Optimum has divided its mainnet offering into two services. Signal accelerates incoming beacon-block delivery and is available free to any Ethereum validator operator, while Accelerate focuses on sending proposed blocks faster and optimizing how long proposers can wait before selecting a builder bid.
Accelerate currently has a narrower admission threshold. Self-service access requires an operator to control at least 200 validator keys, reflecting its focus on professional staking operators and block proposers seeking to optimize execution-layer rewards and MEV bid selection.
The rollout therefore fits into a wider effort to improve Ethereum validator operations without redesigning consensus itself. Work on simplifying institutional validator infrastructure has similarly focused on reducing operational friction, while mump2p addresses another part of the stack by attempting to make communication between distributed nodes faster and more predictable.
That networking layer could become more consequential as Ethereum increases Layer 1 capacity. The network’s Glamsterdam scaling roadmap is preparing Ethereum for larger execution workloads, making propagation efficiency increasingly relevant as block sizes, builder workflows and validator timing constraints evolve.
For now, the mainnet launch establishes that mump2p is operating with live Ethereum stake rather than only in controlled tests. The next measure of significance will be whether the reported latency and attestation gains persist across a larger, more geographically diverse validator set, rather than the amount of stake associated with the initial 16 operators.