Data Availability Layer

In the context of blockchain, data availability refers to the assurance by network participants that the data uploaded by users is faithfully proposed and stored in certain nodes of the network, and that these nodes can prove this property to other participants.

To achieve consensus on data availability, the most straightforward approach, such as methods in early blockchain systems like Bitcoin, requires network participants to download all block data and verify its validity, leading to significant scalability issues: if the blockchain stores large enough data, ordinary users may not be able to afford the cost of processing this data or may not wish to spend a lot of bandwidth on data they are not interested in. As a result, they may be unable or unwilling to participate in verification, leading to centralization issues in the network.

To mitigate this challenge, LazyLedger proposed a new technique that uses random sampling to verify data availability. Technically, Data Availability Sampling (DAS) allows network participants to ensure the full availability of block data without requiring any participant to download it completely. This scheme allows potentially malicious block proposers to encode the content of the block into a commitment 𝜎 and a complete encoded block 𝜋. The commitment 𝜎 is added to the block header and it allows light nodes, which are run by ordinary network participants, to verify the availability of the complete 𝜋 by requesting a few positions in 𝜋 randomly. If a sufficient number of light nodes successfully probe 𝜋, DAS ensures that the data is fully available.

Note that a single light node cannot be certain of the full availability of the data, as it queries only a small portion of the encoded data, therefore DAS still requires a sufficiently large group of light nodes.

Last updated