Danksharding is how Ethereum becomes a truly scalable blockchain, but there are several protocol upgrades required to get there. Proto-Danksharding is an intermediate step along the way. Both aim to make transactions on Layer 2 as cheap as possible for users and should scale Ethereum to >100,000 transactions per second.
Proto-Danksharding, also known as EIP-4844(opens in a new tab), is a way for rollups to add cheaper data to blocks. The name comes from the two researchers who proposed the idea: Protolambda and Dankrad Feist. Right now, rollups are limited in how cheap they can make user transactions by the fact that they post their transactions in
CALLDATA. This is expensive because it is processed by all Ethereum nodes and lives on chain forever, even though rollups only need the data for a short time. Proto-Danksharding introduces data blobs that can be sent and attached to blocks. The data in these blobs is not accessible to the EVM and is automatically deleted after a fixed time period (1-3 months). This means rollups can send their data much more cheaply and pass the savings on to end users in the form of cheaper transactions.
Rollups post the transactions they execute in data blobs. They also post a "commitment" to the data. They do this by fitting a polynomial function to the data. This function can then be evaluated at various points. For example, if we define an extremely simple function
f(x) = 2x-1 then we can evaluate this function for
x = 1,
x = 2,
x = 3 giving the results
1, 3, 5. A prover applies the same function to the data and evaluates it at the same points. If the original data is changed, the function will not be identical, and therefore neither are the values evaluated at each point. In reality, the commitment and proof are more complicated because they are wrapped in cryptographic functions.
KZG stands for Kate-Zaverucha-Goldberg - the names of the three original authors(opens in a new tab) of a scheme that reduces a blob of data down to a small cryptographic "commitment"(opens in a new tab). The blob of data submitted by a rollup has to be verified to ensure the rollup is not misbehaving. This involves a prover re-executing the transactions in the blob to check that the commitment was valid. This is conceptually the same as the way execution clients check the validity of Ethereum transactions on layer 1 using Merkle proofs. KZG is an alternative proof that fits a polynomial equation to the data. The commitment evaluates the polynomial at some secret data points. A prover would fit the same polynomial over the data and evaluate it at the same values, checking that the result is the same. This is a way to verify the data that is compatible with zero-knowledge techniques used by some rollups and eventually other parts of the Ethereum protocol.
A KZG ceremony is a way for many people from across the Ethereum community to generate a secret random string of numbers together that can be used to verify some data. It is very important that this string of numbers is not known and cannot be recreated by anyone. To ensure this, each person that participates in the ceremony receives a string from the previous participant. They then create some new random values (e.g. by allowing their browser to measure the movement of their mouse) and mix it in with the previous value. They then send the value on to the next participant and destroy it from their local machine. As long as one person in the ceremony does this honestly, the final value will be unknowable to an attacker. The EIP-4844 KZG ceremony was open to the public and tens of thousands of people participated to add their own entropy. For the ceremony to be undermined, 100% of those participants would have to be actively dishonest. From the perspective of the participants, if they know they were honest, there is no need to trust anyone else because they know that they secured the ceremony (they individually satisfied the 1-out-of-N honest participant requirement).
Danksharding is the full realization of the rollup scaling that began with Proto-Danksharding. Danksharding will bring massive amounts of space on Ethereum for rollups to dump their compressed transaction data. This means Ethereum will be able to support hundreds of individual rollups with ease and make millions of transactions per second a reality.
The way this works is by expanding the blobs attached to blocks from 1 in Proto-Danksharding to 64 in full Danksharding. The rest of the changes required are all updates to the way consensus clients operate to enable them to handle the new large blobs. Several of these changes are already on the roadmap for other purposes independent of Danksharding. For example, Danksharding requires proposer-builder separation to have been implemented. This is an upgrade that separates the tasks of building blocks and proposing blocks across different validators. Similarly, data availability sampling is required for Danksharding, but it is also required for the development of very lightweight clients that do not store much historical data ("stateless clients").
Full Danksharding is several years away. However, Proto-Danksharding should arrive relatively soon. At the time of writing (Feb 2023) the KZG ceremony is still open and has so far attracted over 50,000 contributors. The EIP(opens in a new tab) for Proto-Danksharding is mature, the specification is agreed and the clients have implemented prototypes that are currently being tested and made production-ready. The next step is to implement the changes on a public testnet. You can keep up to date using the EIP 4844 readiness checklist(opens in a new tab).
- Proto-Danksharding notes(opens in a new tab) - Vitalik Buterin
- Dankrad's notes on Danksharding(opens in a new tab)
- Dankrad, Proto and Vitalik discuss Danksharding(opens in a new tab)
- The KZG ceremony(opens in a new tab)
- Carl Beekhuizen's Devcon talk on trusted setups(opens in a new tab)
- More on data availability sampling for blobs(opens in a new tab)
- Dankrad Feist on KZG commitments and proofs(opens in a new tab)
- KZG polynomial commitments(opens in a new tab)