Why Modular Data Availability Is the Next Frontier for AVS Security

On monolithic chains like Ethereum, L2s spend >90% of transaction fees on data posting. This creates a direct conflict: scaling throughput exponentially increases the DA bill, making cheap transactions impossible.

• Economic Limit: A 100 TPS L2 can incur >$1M/month in pure Ethereum calldata costs.
• Security Trade-off: To cut costs, chains are forced to consider less secure DA layers, creating systemic risk.

Each new AVS (e.g., EigenLayer, Babylon) requires its own set of attestations—proofs of correct state—to be published. A monolithic DA layer becomes a single point of congestion and failure for hundreds of security services.

• Data Bloat: A network of 50 AVSs could require >1 TB/day of attestation data.
• Fragile Security: Congestion on the base layer delays critical fraud proofs, extending the window for malicious attacks.

Secure bridging and shared sequencing between rollups (e.g., using LayerZero, Hyperlane) require verifiable access to each other's state. Relying on different, opaque DA layers makes this cryptographically impossible without trusted committees.

• Verification Gap: A ZK-proof for a cross-rollup message must have a cryptographic root in a universally available data layer.
• Fragmented Liquidity: Without a shared DA standard, interoperability reverts to the multi-sig and trusted oracle model, the very weakness DeFi aims to solve.

Celestia decouples consensus and execution, offering a pluggable DA layer that scales with the number of users. Its core innovation is Data Availability Sampling (DAS), allowing light nodes to securely verify data availability without downloading entire blocks.

• Enables sovereign rollups with full control over their stack.
• Linear scaling; cost per byte decreases as more rollups join the network.
• Foundation for modular stacks like Arbitrum Orbit and OP Stack.

Built by EigenLayer, EigenDA leverages restaked ETH to secure data availability. It turns Ethereum's economic security into a reusable commodity for AVSs, creating a cryptoeconomic flywheel.

• Native Ethereum security via restaking, avoiding new trust assumptions.
• High throughput design targeting 10-100 MB/s blob capacity.
• Integrated stack for AVSs using EigenLayer for both security and DA.

Avail provides a scalable DA layer built with a Polkadot SDK-based Proof-of-Stake consensus, focusing on verifiability and seamless Ethereum integration through bridges.

• Focus on light client verifiability with validity and KZG commitment proofs.
• Ethereum interoperability via a secure bridge for unified liquidity.
• Modular tooling including Avail Nexus for cross-chain unification.

NEAR Protocol's DA solution uses its existing Nightshade sharding architecture to offer high-throughput, low-cost data posting. It's a battle-tested layer repurposed for modular chains.

• Sharded architecture provides inherent horizontal scalability.
• Proven production use securing the NEAR L1 and ~$200M+ in rollup contracts.
• Cost-competitive pricing, often ~100x cheaper than Ethereum calldata.

Using Ethereum mainnet for DA forces rollups to pay premium gas fees for calldata, which doesn't scale. This creates a security vs. cost trade-off that stifles adoption.

• EIP-4844 (blobs) are a stopgap, not a long-term scaling solution.
• High fixed cost for security limits experimentation for new AVSs.
• Monolithic thinking in a modular world creates economic inefficiency.

Modular DA transforms availability from a bottleneck into a competitive, commoditized service. This lets AVS architects optimize for specific trade-offs: cost, latency, and security source.

• Unlocks vertical integration (e.g., a rollup using EigenDA for shared security).
• Drives cost discovery through market competition between Celestia, EigenDA, Avail.
• Future-proofs security by separating DA from execution, enabling rapid iteration.

A malicious sequencer or DA layer can withhold transaction data, preventing fraud proofs from being constructed. This turns a liveness failure into a safety failure, allowing invalid state transitions to finalize.

• Attack Vector: Sequencer posts only state root to L1, hides data.
• Critical Dependency: Relies on data availability sampling (DAS) and fishermen for security.
• Systemic Risk: Can affect all rollups using that DA layer (e.g., Celestia, EigenDA, Avail).

DA layers become centralized profit centers. A dominant provider like Celestia or EigenDA can censor specific rollups or extract maximal value via MEV-aware data ordering, undermining neutrality.

• Economic Power: Control over ~$10B+ in sequencer fees.
• Trust Assumption: Rollups must trust the DA layer's governance and economic incentives.
• Mitigation: Requires multi-DA clients and proof-of-censorship protocols.

AVSs like EigenLayer restakers secure multiple modular systems. A correlated slashing event on a faulty DA layer can cause cascading insolvency across unrelated AVSs, violating risk isolation.

• Correlation Failure: A single DA fault triggers slashing across hundreds of AVSs.
• Capital Efficiency Trap: High restaking yields mask aggregated tail risk.
• Solution: Requires risk-tiered slashing and explicit AVS-to-DA dependency graphs.

DA layers rely on light clients for bridge security (e.g., IBC, layerzero). A successful 51% attack on the DA layer can forge fraudulent state proofs, draining all connected bridges and rollups.

• Verification Gap: Light clients assume 1/N trust in validator sets.
• Bridge Target: Wormhole, Axelar, and Across are high-value targets.
• Requirement: ZK-proofs of DA (e.g., zkPorter, Near DA) are the endgame.

Relying on a single chain (e.g., Ethereum L1) for data availability creates a critical vulnerability and cost center.\n- Security Ceiling: AVS security is capped by the DA layer's consensus.\n- Cost Inefficiency: Paying for global consensus when you need localized verification.\n- Throughput Bottleneck: All AVS blobs compete for the same scarce block space.

Specialized DA layers decouple execution from consensus and availability, enabling scalable security.\n- Data Availability Sampling (DAS): Light clients can verify terabytes of data with minimal resources.\n- Economic Security Stacking: AVS can leverage restaked ETH (EigenDA) or a dedicated token (Celestia) for crypto-economic security.\n- Interoperable Standards: Blobstream proofs bridge DA guarantees back to L1s like Ethereum and Arbitrum.

Modular DA transforms security from a fixed cost to a configurable variable, enabling new trust models.\n- Security-as-a-Service: Procure DA from the most cost-effective provider (Celestia, EigenDA, Avail) for your risk profile.\n- Multi-Homing AVS: Distribute data across multiple DA layers for censorship resistance.\n- Verifiable Off-Chain DA: Use validity proofs (zk-proofs) to attest to data availability, enabling high-throughput, low-cost L2s.

Isolated DA is useless; the value is in proving data availability across ecosystems.\n- Universal Settlement: Ethereum L1 remains the root of trust via bridges like EigenDA's proof system or Celestia's Blobstream.\n- Cross-Rollup Composability: Shared DA enables native, trust-minimized communication between rollups (e.g., using zk-proofs of state transitions).\n- Avoiding Vendor Lock-In: Standards like EIP-4844 blobs and Celestia's Namespaced Merkle Trees ensure AVS aren't trapped in one stack.