Data Availability Sampling (DAS) is a protocol that lets nodes verify a block's data is fully available without downloading the entire block. Instead of pulling every piece of data, a node randomly samples a small number of fragments and checks their integrity. If the samples verify correctly, statistical guarantees allow the node to infer that the full block is present and correct. Modern blockchains targeting high throughput generate large blocks, often several megabytes. Without DAS, every validator would need to download and store the full block, creating a bottleneck that limits participation and raises costs. DAS keeps validation lightweight, allowing nodes on modest hardware to participate, which preserves decentralization. DAS also defends against data-withholding attacks. If an adversary hides part of a block, randomly sampling nodes will likely hit the gap and reject the block. This makes DAS essential for rollup architectures that rely on cheap data availability proofs, and for Ethereum's danksharding roadmap.
Back to Glossary