Data Availability Committee (DAC)

A Data Availability Committee (DAC) is a permissioned set of known, often reputable, nodes that cryptographically sign attestations guaranteeing that the data for a rollup or state channel is available for download. This model is a cornerstone of validium and certain volition architectures, where transaction data is stored off-chain. By removing the requirement to post all data to the base layer (e.g., Ethereum), these systems achieve significant scalability and cost reductions. The security assumption shifts from the base layer's consensus to the trustworthiness and liveness of the committee members.

The core function of a DAC is to provide a data availability attestation. When a rollup operator publishes a state root or zero-knowledge proof to the main chain, they also submit signatures from a supermajority of the DAC members. These signatures assert that the corresponding batch data is available on the DAC's servers. Any user or light client can then challenge the system by requesting this data; if it is not provided, the attestation is proven false, preventing fraudulent state transitions. This creates a cryptographic safety net reliant on the committee's honesty.

Compared to other data availability solutions like data availability sampling (DAS) on danksharding, a DAC represents a more centralized, trust-optimized approach. Its security is cryptoeconomic, backed by the reputational stake and legal identity of its members, which may include foundations, exchanges, or other institutional players. Prominent examples include StarkEx's DAC for its Validium instances, where members like Nethermind and ConsenSys serve on the committee. This design makes DACs suitable for high-throughput, enterprise-grade applications where absolute decentralization is traded for performance and cost.

A Data Availability Committee (DAC) is a permissioned group of reputable, independent entities—often foundations, exchanges, or institutional validators—that collectively sign cryptographic attestations, or signatures, confirming they possess and are storing the complete data for a block. This attestation is published on the parent chain (like Ethereum) as a compact proof. Instead of every network participant downloading all data to verify its availability, they can trust the committee's signed commitment. This model significantly reduces the data verification burden, enabling higher throughput and lower costs for Layer 2 rollups and validiums, which do not post their full transaction data directly to the base layer.

The operational workflow involves several key steps. When a rollup's sequencer produces a new batch of transactions, it sends the associated data to each DAC member off-chain. Each member independently verifies the data's integrity and completeness. Upon successful verification, they cryptographically sign a data availability attestation, typically a Merkle root hash of the data. A threshold of signatures (e.g., a majority or supermajority) is aggregated into a single multi-signature, which is then posted to the main chain. This attestation acts as a bonded promise; if the data is later found to be unavailable, the committee members' staked collateral can be slashed.

This trust model presents a clear security trade-off. A DAC offers robust liveness guarantees and high performance but introduces a weaker security assumption than data availability sampling (DAS) used in solutions like celestia or Ethereum danksharding. The system's security depends on the honesty and coordination of the committee members. If a malicious actor compromises a threshold of members, they could collude to sign for unavailable data, potentially enabling fraud that users cannot detect or challenge. Therefore, the credibility, decentralization, and economic stake of the DAC members are paramount to the system's overall security.

DACs are a pragmatic solution for applications prioritizing extreme scalability and cost-efficiency where absolute cryptographic security is a secondary concern. They are commonly employed in validium architectures, such as those used by StarkEx, and certain zk-Rollup configurations. In these systems, validity proofs (ZK-SNARKs/STARKs) ensure transaction correctness, while the DAC ensures data liveness. This hybrid approach—cryptographic guarantees for execution and a trusted committee for data—provides a powerful balance for enterprise-grade or high-frequency trading applications where data can be managed off-chain with reputable custodians.

The evolution of data availability solutions is moving towards trust-minimized designs. While DACs serve as a critical bridging technology, many projects view them as an interim step toward fully cryptoeconomic or cryptographic data availability. Future systems aim to replace the trusted committee with a large, permissionless network of nodes performing data availability sampling, thereby removing the need for trust in a specific set of entities and aligning more closely with blockchain's core decentralization ethos.

A Data Availability Committee (DAC) is a trusted, permissioned group of entities that cryptographically attest to the availability of transaction data for a Layer 2 (L2) rollup, serving as a simpler, more centralized alternative to full on-chain data posting. In this model, the rollup's sequencer sends data to committee members who sign a commitment, typically a Merkle root, which is then posted to the parent chain (Layer 1). Users must trust that at least one honest committee member will provide the data if challenged, as the L1 contract only verifies the aggregate signature, not the data itself. This design significantly reduces transaction costs compared to posting all data on-chain but introduces a trust assumption regarding the committee's honesty and liveness.

The evolution from DACs represents a shift toward trust-minimized and cryptoeconomically secure systems. The primary successor architecture is Data Availability Sampling (DAS), as implemented by networks like Celestia and Ethereum's danksharding roadmap. DAS replaces the trusted committee with a network of light nodes that perform multiple rounds of random sampling on erasure-coded data. If the data is unavailable, sampling will eventually fail with high probability, providing probabilistic security without needing to download the entire dataset. This creates a scalable data availability layer that is secure under an honest minority assumption, a significant reduction in trust compared to DACs.

The final stage in this progression is the move to full on-chain data availability, where rollup transaction data is posted directly and entirely to a Layer 1 blockchain, such as Ethereum's calldata or blobs. This provides the highest level of security, inheriting the full consensus-level security and censorship resistance of the underlying L1. The trade-off is cost, as storing data on a high-security chain is expensive. Innovations like EIP-4844 (proto-danksharding) with blob-carrying transactions are designed to reduce this cost while maintaining this strong security model, making pure DAC architectures largely obsolete for rollups prioritizing decentralization.