A Protocol Design View on Statelessness

Julian Ma

Who Should Hold State?

Statelessness is about choosing Proof-Serving Node architecture and State Growth jointly.

Attesters should not be the proof-serving nodes, but who should be it?

1. Assuming a weak stateless protocol, is it necessary to implement partial statelessness in protocol?

​

• What is the fundamental difference between weak and strong statelessness?

​

• How does state growth constrain the gas limit in a weak stateless protocol?

Out-of-protocol Partial Statelessness

​

• Partial statelessness uses a state expiry rule to partition state into active and inactive sets.
• Attesters should all adhere to the same state expiry rule.
• Block producers, can determine one themselves!

Why would block producers do so?

1. Some state is not economically meaningful.

Similarities to the MEV-Boost market: builders leave out transactions because their value is negligible.

• State growth too large.

Positive framing: if state growth is too large, builders can prune state tree without in protocol state expiry.

​

Difference Weak and Strong Statelessness

• Does the block producer (weak) or user (strong) store the state?

​

• Difference is honest block producer behavior.

​

• Cannot rely on honest block producers, should find other parties to store the state.

​

• Honest block producer definition is still important for FOCIL: do transactions need a witness or not.

State Growth after Statelessness

• Who holds the state?
• Only one non-censoring person needed, but it should be clear who this person is.

​

• State growth set such that they can reliably provide up-to-date witness to users.

​

State Growth after Statelessness

• Who holds the state?
• Only one non-censoring person needed, but it should be clear who this person is.

​

• State growth set such that they can reliably provide up-to-date witness to users.

​

dApps and wallets?

Thanks!

Telegram: @julianzma