General thoughts on essays by Vitalik about Ethereum and blockchain

September 12, 2022

• the merge

  • the consensus layer drives network efficiency -> moving to POS
  • the execution layer drives network performance and scalability

• https://vitalik.ca/general/2021/12/06/endgame.html - how to keep large

  • blockchains acceptably trustless and secure
  • different staking tiers, fraud proofs, zk-snarks (basically move a lot of work from mainnet to secondary channels)
  • block production remains centralized, block validation becomes more trustless and decentralised

• https://vitalik.ca/general/2021/05/23/scaling.html - limits to scalability

  • imperative that regular users can run nodes for decentralisation
  • considerations around compute, bandwidth, storage and associated interaction effects
  • for ethereum, medium-term solutions include sharding (split network horizontally to spread load), statelessness and state expiry (validation without permanent storage), zk-snarks (perform transactions / validations / cryptographic proofs off-chain, compress this information, commit back on-chain)

• https://vitalik.ca/general/2020/08/17/philosophy.html - blockchain validation

  • network designs and the role of being a validator
  • general thoughts are around light clients and sharding in the context of validation (shards introduce more transactions to validate, everyone becomes a light client, fraud proofs may highlight and communicate invalid blocks, each shard requires a minimum threshold of participation for herd immunity)
  • defaulting to victory vs defaulting to chaos (the latter making it very expensive to run 51 percent attacks ie losing stakes and ensuring immutability above all else, otherwise we will just rewrite history / fork the network)