1. Simplify Ethereum node architecture. The EL can become a library instead of being a full blown node.
2. Reduce sync time. A node can sync by only downloading the beacon state. This can enable syncing in a few seconds.
Running a node as part of your wallet software (even on your mobile phone) becomes practical.
3. Improve Ethereum decentralization.
The most important parameter for decentralization is the cost of running a full node. Verkle reduces it both operationally (see 1 and 2) and in terms of hardware cost.
4. Enable the way for full blown light nodes.
In the endgame light nodes should be fully protected against invalid state transitions. In a world where zk proving does not achieve this for consensus+execution layer, verkle enables simple fraud proofs for execution payloads.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
I'm thinking about a new sharding design, where instead of having independent proposers for each shard, all shard blocks in one slot are proposed together with the beacon block. This leads to a major simplification of the sharding design 1/n
In the previous design, at each slot, shard blocks are independently proposed and data availability has to be verified by committees. We can't just do that check globally for all shards or each proposer has the ability to disrupt the whole process (liveness failure) 2/n
In most cases, the next beacon block would have info on all the shard data that was confirmed, but this is not guaranteed: Voting might take longer, especially if shard proposers intentionally make data marginally available. This makes things quite complex. 3/n