<aside>
☝ Data Availability Sampling (DAS) is a key primitive to scaling blob space for L2s on Ethereum. The first step is PeerDAS – scheduled to ship in the Fusaka hard fork. The focus of this site for now is on PeerDAS, but will expand over time.
DAS.atm
</aside>
https://www.youtube.com/watch?v=WOdpO1tH_Us
💡 WTF is Data Availability Sampling?
- A cryptographic technique that allows nodes to only download a few small chunks of data (samples) to verify that all data has been made available (block is valid).
💡 WTF is PeerDAS?
- Proposal to extend the current beacon chain p2p network to support sharing data samples instead of entire block data. This is an intermediate step on the way to full DAS.
<aside>
🏠
Handy Links
</aside>
DAS what, why, how?
who.DAS?
Roadmap
Other Writings
Videos
<aside>
📌
Resources
</aside>
<aside>
📜
History of Blobs
</aside>
<aside>
1️⃣
WTF are…Blobs? (EIP-4844, Dencun)
- Blobs: Data chunks from EIP-4844 let L2 rollups post tx data to L1 cheaply.
- At 3 target/6 max blobs per block, they cut fees, boost capacity, and prune after 18 days.
- Data availability ensures trust by proving data isn’t hidden and verifiable.
</aside>
<aside>
2️⃣
WTF is…Blob Scaling? (EIP-7691, Pectra)
- EIP-7691 ups blob count to 6/9 per block, doubling throughput for L2s.
- Post-Dencun tests show network stability and bandwidth available to justify a short-term scale boost prior to PeerDAS.
- Blob fees drop ~14.5% when empty, rise ~8.2% when full, a 2:3 target/max ratio.
</aside>
<aside>
3️⃣
WTF is…PeerDAS? (EIP-7594, Fusaka)
- PeerDAS scales to 12 blobs/block via erasure coding and sampling—full nodes and supernodes check data chunks, not all.
- P2P sharing keeps full nodes light, supernodes robust, boosting L2 capacity.
WTF is…a Full Node in PeerDAS?
- Full Nodes: Regular Ethereum nodes that custody a small slice (e.g., 4 columns) of blob data, sample 8 random columns per slot, and verify availability via p2p gossip. They keep the network light and decentralized, using modest bandwidth like today’s mainnet.
WTF is…a Supernode in PeerDAS?
- Supernodes: Beefy nodes (often run by pros) that custody all blob columns, sample everything, and help propagate data fast with distributed blob building. They require more CPU and bandwidth to ensure blobs scale without hiccups.
Other Writings on PeerDAS
</aside>
<aside>
👀
Follow the latest / find ways to help
</aside>
For more EF Research Updates, check out:
<aside>
<img src="/icons/arrow-right-basic_lightgray.svg" alt="/icons/arrow-right-basic_lightgray.svg" width="40px" />
EF Research
</aside>