What is Data Availability? Quoting Paradigm’s blog here
Somebody (such as a layer-1 block proposer or layer-2 sequencer) has produced a block of data. They claim to have made it “available” to the “public.” Your goal is to check the availability claim, i.e., would you actually be able to obtain the data if you needed to?
Availability of data is crucial. Optimistic fraud-proof-based systems, like Optimism, require data availability for verification, and even validity-proof-based systems, like StarkNet or Aztec, require data availability to ensure liveness (e.g., to prove asset ownership for a rollup’s escape hatch or forced transaction inclusion mechanism).
Data Availability Sampling Requirements
DAS is based on -
**Erasure coding** - is an important technique that makes data availability sampling possible. Erasure coding extends a block by producing additional copies of the data. The additional data creates redundancy, giving stronger security guarantees for the sampling process.
Proof of correct encoding - Nodes may try to incorrectly encode data to disrupt the network. To defend against such an attack, we need a way to verify the correctness of the encoding - this is where the correctness proofs come in. Proofs can be commitment proofs based on Zero knowledge (KZG based, used in Eigen DA, Avail and Ethereum Danksharding) or Fraud proofs (used in Celestia).
A typical DAS process, this is described in Eth context but process remains the same more or less.