
Briefing
Existing Data Availability Sampling (DAS) relies on fixed-rate erasure codes, restricting light nodes to sampling from a predetermined set of symbols, which limits the expressive power of the samples and the cryptographic assurance of data availability. This paper proposes a new paradigm that commits to the uncoded data and employs on-the-fly coding via Random Linear Network Coding (RLNC). This fundamental shift in the cryptographic primitive enables light nodes to request random linear combinations of the data, thereby achieving significantly stronger assurance of data availability and solidifying the security foundation for all high-throughput, modular blockchain architectures.

Context
The foundational problem in modular blockchain design is ensuring data availability without requiring all nodes to download the entire block, a challenge addressed by Data Availability Sampling (DAS). Established DAS schemes, such as those utilizing Reed-Solomon or low-density parity-check codes, require the block producer to commit to a fixed, pre-coded block. This model forces light nodes to sample from a limited, predetermined set of coded symbols, which restricts the strength of the security guarantee and creates a theoretical limitation on the confidence level achievable by a light node.

Analysis
The core mechanism introduces a modularization of the data availability process by committing to the uncoded data directly. This new model then employs Random Linear Network Coding (RLNC) for on-the-fly sampling. The fundamental difference lies in the commitment ∞ previous schemes commit to the erasure-coded block, whereas this scheme commits to the raw data.
When a light node samples, it requests a random linear combination of the data, which is generated by the network on-the-fly using RLNC. This ensures that the block producer must make the original data available for any random linear combination to be verifiable, as any linear combination of the original data is equally informative, providing a more robust and expressive sampling space than fixed-rate symbol sampling.

Parameters
- Assurance Improvement ∞ Multiple orders of magnitude stronger assurances ∞ The new paradigm provides a significantly higher security guarantee of data availability compared to sampling pre-committed symbols.

Outlook
This research establishes a new, more robust cryptographic primitive for data availability, which is a critical building block for the next generation of modular systems. The immediate next step involves formalizing the integration of this RLNC-based DAS into existing rollup frameworks and optimizing the commitment scheme for practical deployment overhead. In the 3-5 year horizon, this theory could unlock truly ubiquitous and trust-minimized light client participation, fundamentally enabling a new level of scalability and decentralization by removing the last vestiges of reliance on fixed, less expressive cryptographic sampling.
