Modular Blockchain
A blockchain architecture separating execution, consensus, data availability, and settlement into specialized layers for scalability.
Key Takeaways
- A modular blockchain separates its core functions into specialized layers: execution, settlement, consensus, and data availability. Each layer can be independently optimized, unlike monolithic chains that handle everything on a single network.
- The modular stack enables horizontal scaling: dedicated data availability layers like Celestia handle storage, rollups handle execution, and base chains like Bitcoin or Ethereum provide final settlement.
- Bitcoin is increasingly positioned as a modular settlement layer through technologies like BitVM and ZK rollups, allowing it to anchor off-chain computation without protocol changes.
What Is a Modular Blockchain?
A modular blockchain is a blockchain architecture that delegates its core functions to separate, specialized layers rather than handling everything on a single chain. Instead of one network processing transactions, ordering them, storing their data, and finalizing their results, a modular design splits these responsibilities across purpose-built systems that coordinate through cryptographic proofs.
The concept draws from the software engineering principle of separation of concerns. A monolithic blockchain like the original Ethereum design or Solana performs all four core functions within a single validator set. A modular architecture lets each layer optimize for its specific task: an execution layer can maximize throughput without worrying about data storage, while a data availability layer can scale storage without needing to process transactions.
The term "modular blockchain" was introduced in a 2018 academic paper co-authored by Mustafa Al-Bassam and Vitalik Buterin on fraud and data availability proofs. Al-Bassam later expanded the concept in his 2019 LazyLedger whitepaper, which proposed a chain that functions purely as a data availability layer. LazyLedger eventually became Celestia, the first dedicated modular data availability network.
How It Works
Every blockchain, whether monolithic or modular, must perform four core functions. In a modular architecture, each function is handled by a dedicated layer:
The Four Core Functions
Execution is where transactions and smart contract logic are actually computed. Account states update, balances change, and contract code runs. In a modular stack, execution is typically handled by rollups that process transactions off-chain and submit compressed results to other layers.
Settlement is the layer where transaction validity is established and finalized. It serves as the ultimate source of truth, handling dispute resolution through fraud proofs (optimistic rollups) or validity proofs (ZK rollups). Bitcoin and Ethereum both serve as settlement layers in modular designs.
Consensus determines the ordering of transactions and ensures all participants share an identical view of the ledger. In a modular system, consensus can be handled by a dedicated layer separate from execution, allowing the ordering mechanism to be optimized independently.
Data availability guarantees that all data necessary for validating transactions is accessible to network participants. Without data availability, validators cannot verify that the execution layer computed correctly. This is the function that projects like Celestia, EigenDA, and Avail specialize in.
Data Availability Sampling
A key innovation enabling modular blockchains is data availability sampling (DAS). Instead of requiring every node to download all block data, DAS allows light nodes to verify data availability by randomly sampling small portions of block data. The technique uses 2D Reed-Solomon coding to split data into chunks encoded in a matrix, then uses namespaced Merkle trees for organized retrieval.
With each additional sample, confidence that all data is available increases exponentially. This means nodes with minimal hardware can still verify data availability, keeping the network decentralized even as data throughput scales.
How Layers Coordinate
A typical modular stack operates in the following sequence:
- Users submit transactions to a rollup (execution layer)
- The rollup orders and executes transactions, updating state
- Transaction data is posted to a data availability layer (Celestia, Ethereum blobs, or on-chain)
- The rollup submits a proof (fraud proof or validity proof) to the settlement layer
- The settlement layer verifies the proof and finalizes the state transition
Each step involves a different specialized system, but they are cryptographically linked. A fraud proof on the settlement layer references data on the DA layer, which in turn corresponds to execution on the rollup. If any layer behaves dishonestly, the proofs will fail.
Modular Stack (example configuration)
┌─────────────────────────────────┐
│ Execution Layer (Rollup) │ ← processes transactions
│ e.g., Arbitrum, Citrea │
└──────────┬──────────────────────┘
│ posts data
┌──────────▼──────────────────────┐
│ Data Availability Layer │ ← stores transaction data
│ e.g., Celestia, EigenDA │
└──────────┬──────────────────────┘
│ submits proofs
┌──────────▼──────────────────────┐
│ Settlement Layer │ ← verifies proofs, finalizes state
│ e.g., Ethereum, Bitcoin │
└─────────────────────────────────┘The Modular Ecosystem
Data Availability Layers
Celestia is the most prominent dedicated DA layer, commanding roughly half the DA market as of 2026. Its Matcha upgrade in November 2025 expanded block size from 8MB to 128MB, and over 56 rollups use Celestia for data availability across mainnet and testnet. EigenDA, built on EigenLayer's restaking protocol, achieves 100 MB/s throughput on Ethereum mainnet and is backed by over $19 billion in restaked ETH. Avail takes an ecosystem-agnostic approach with its three-layer stack: Avail DA, Nexus (unification), and Fusion (additive security).
Execution Layers
Rollups serve as the primary execution layers in a modular stack. Optimistic rollups like Arbitrum and Optimism assume transactions are valid unless challenged via fraud proofs. ZK rollups like zkSync and StarkNet generate cryptographic zero-knowledge proofs that mathematically verify validity without revealing full transaction details. Each approach presents different tradeoffs between proof generation cost, finality speed, and EVM compatibility.
Settlement Layers
Ethereum has explicitly adopted a rollup-centric roadmap, with its Dencun upgrade in March 2024 introducing EIP-4844 (Proto-Danksharding). This created blob-carrying transactions: a new transaction type where rollups attach temporary data chunks to the consensus layer, cutting data posting costs by 10 to 100x. Subsequent upgrades (Pectra in May 2025, Fusaka in December 2025) have continued expanding blob capacity.
Bitcoin is also being used as a settlement layer, particularly through projects like Citrea (Bitcoin's first production ZK rollup, launched January 2026) and BitVM, which enables optimistic verification on Bitcoin without protocol changes. Unlike Ethereum's explicit modular support, Bitcoin's role as a settlement layer works within its existing constraints, aligning with its conservative design philosophy.
Modular vs. Monolithic
Understanding modular blockchains requires comparing them against the monolithic alternative. Monolithic chains like Solana and the original Ethereum design handle all four functions within a single validator set.
| Aspect | Monolithic | Modular |
|---|---|---|
| Scalability | Limited by single-layer capacity | Horizontal scaling through parallel layers |
| Security model | Unified validator set, simpler trust model | Multiple trust assumptions across layers |
| Composability | Applications interact seamlessly on one ledger | Cross-rollup composability remains challenging |
| Customizability | Constrained by network-wide rules | Choose rollup framework, DA layer, settlement |
| Complexity | Simpler to deploy and audit | More moving parts and integration surfaces |
| Decentralization | Hardware requirements grow with throughput | Light nodes verify via DAS with lower barriers |
The industry consensus as of 2026 is that both approaches will coexist. Modular architectures lead in TVL growth and developer activity, while monolithic chains remain competitive on raw throughput and user experience simplicity. Independent benchmarks from 2025 showed modular architectures delivering over 6x higher throughput at significantly lower cost than equivalent monolithic setups.
Why It Matters
The modular thesis has practical implications for how blockchain networks scale. Rather than requiring every validator to process every transaction, modular architectures let networks grow by adding specialized capacity. A rollup can scale execution independently of the settlement layer's throughput limits.
For Bitcoin specifically, the modular approach is significant because it allows the base layer to remain conservative and secure while enabling high-throughput applications on top. Layer 2 protocols built on Bitcoin can handle thousands of transactions per second without changing Bitcoin's core protocol. Projects like Spark demonstrate this approach: by building an off-chain protocol anchored to Bitcoin's settlement guarantees, Spark enables instant, feeless BTC transfers while inheriting Bitcoin's security. For a broader comparison of Bitcoin scaling solutions, see the Bitcoin Layer 2 comparison.
Use Cases
- Scalable rollup deployments: application-specific rollups can be launched with custom execution environments while sharing a common settlement and data availability layer, reducing the cost and complexity of launching a new chain
- Cost-efficient data storage: by separating data availability from execution, rollups can post transaction data to specialized DA layers at a fraction of the cost of storing it directly on a settlement chain like Ethereum
- Sovereign chains: projects can deploy sovereign rollups that use an external DA layer (like Celestia) but maintain their own settlement rules, giving them full control over governance and upgrade paths
- Bitcoin-anchored applications: using Bitcoin as the settlement layer, ZK rollups and other L2 protocols can build complex applications anchored to Bitcoin's security without requiring protocol changes
- Cross-chain interoperability: modular designs enable shared data availability layers to serve as coordination points between different execution environments, simplifying cross-chain communication
Risks and Considerations
Security Fragmentation
A monolithic chain has a single security model: one validator set, one consensus mechanism, one set of trust assumptions. A modular stack introduces multiple trust boundaries. The execution layer, DA layer, and settlement layer each have their own security properties. A vulnerability in any layer can compromise the entire stack. The overall security is only as strong as the weakest link.
Liquidity Fragmentation
When execution is split across many rollups, liquidity fragments across them. Users and applications on different rollups cannot interact directly without bridges, which introduce additional trust assumptions and latency. This contrasts with monolithic chains where all applications share a single liquidity pool and can compose atomically.
Complexity and User Experience
Modular architectures are inherently more complex than monolithic designs. Users may need to manage assets across multiple layers, understand different finality guarantees, and navigate bridging between rollups. Chain abstraction middleware is emerging to hide this complexity, but the technology is still maturing.
Data Availability Guarantees
External DA layers are relatively new infrastructure. If a DA layer fails or withholds data, rollups relying on it could be unable to prove their state, potentially freezing user funds. While DAS provides strong probabilistic guarantees, the long-term reliability of dedicated DA networks under adversarial conditions remains to be proven at scale.
This glossary entry is for informational purposes only and does not constitute financial or investment advice. Always do your own research before using any protocol or technology.