DAS

data availability sampling is a technique that allows light nodes to verify that a block's data is available without downloading the entire block. the cost per verifier is O(sqrt(n)) instead of O(n), where n is the block data size

mechanism

the block producer arranges data in a 2D matrix and applies Reed-Solomon erasure coding along both rows and columns. this extends a k x k data matrix into a 2k x 2k coded matrix. the producer commits to each row and column using a NMT (namespaced Merkle tree) or a polynomial commitment

light nodes sample random cells from the extended matrix and request the corresponding data and Merkle proofs. if a malicious producer has withheld any portion of the data, the erasure coding guarantees that the withholding is detectable with high probability after O(sqrt(n)) samples

properties

the 2D Reed-Solomon structure ensures that if at least 50% of the extended data is available, the entire original data can be reconstructed. combined with random sampling, this gives exponential confidence: k random samples catch a withholding attack with probability 1 - (1/2)^k

role in cyber

BBG uses DAS to ensure data availability across the network. validators and light nodes sample encoded chunks rather than downloading full blocks. structural sync coordinates the sampling protocol — each participant verifies local samples and contributes to the collective availability guarantee

see erasure coding for the Reed-Solomon foundation. see NMT for the commitment scheme. see BBG for the data availability layer. see structural sync for the sync protocol