How Does EigenDA Work?

Erasure-coded blob distribution across restaked operator set

EigenDA splits data blobs into erasure-coded chunks distributed across operators. Each operator stores only a fraction of each blob, reducing per-operator cost while maintaining data reconstructibility.

KZG polynomial commitments and proofs for data correctness verification

Disperser generates KZG commitments and per-chunk proofs certifying data correctness. Operators verify their chunk against the commitment before attesting to storage.

Centralized disperser encoding and distributing data to operator nodes

The disperser encodes data, generates proofs, and distributes chunks to operators. Currently a centralized component — a single point of censorship and availability risk.

EigenLayer AVS with restaked ETH securing data availability guarantees

EigenDA operates as an AVS on EigenLayer, with 4.3M ETH staked as economic security. Operators opt in with delegated restaked ETH and are subject to slashing for data unavailability.

Operator signature attestation for data chunk storage

Operators generate signatures certifying chunk storage. Aggregated signatures are submitted on-chain to certify data availability for a blob.

Retrieval nodes collecting chunks from operators to reconstruct original data

Retrieval nodes collect data shards from validator operators and reconstruct the original data using erasure coding. Data reconstruction requires only a subset of all chunks.

100 MB/s throughput with roadmap to hundreds of MB/s and sub-second latency

Current throughput of 100 MB/s significantly exceeds Ethereum's 8.2 MB per block. Roadmap targets further scaling and latency reduction to under one second.

The disperser is currently a centralized component. A compromised or censoring disperser can selectively withhold data from operators, making entire blobs unavailable despite the operator set being honest and ready.

Operators staking on EigenDA share restaked ETH with other AVSs. A slashing event on EigenDA reduces security for other AVSs, and vice versa — creating bilateral contagion risk through shared stake.

Operators may attest to storing chunks without actually retaining the data (lazy operator attack). Detection requires full data reconstruction, which may not happen until data is actually needed by a rollup.

Blob registrations on Ethereum create on-chain costs that scale with blob count. During periods of Ethereum congestion, registration costs could make EigenDA economically unviable for low-value rollups.

Scaling to hundreds of MB/s increases operator bandwidth and storage requirements, potentially centralizing the operator set to well-capitalized entities and reducing the decentralization that underpins security.