opt: [BREAKING] use twox_64 for bucket hashing

[rphmeier]

A seeded twox is suitable for this purpose - each node will have its own map. While it may be possible to convince publicly-accessible RPC nodes to leak information that enables a HashDoS, it seems unlikely for sequencers to leak it.

Furthermore, the complexity of such an attack is limited by the fact that the inputs are Page IDs. In order to write to a particular page ID, the attacker would first need to create accounts (against a cryptographic hash function) that land in those pages.

Computing the blake3 hash of each written bucket was quite CPU-intensive, and so using twox ought to be much faster.

[rphmeier]

• opt: [BREAKING] use twox_64 for bucket hashing #788 👈 (View in Graphite)
• refactor: define hasher traits in nomt-core hasher module and feature-gate #787 : 1 other dependent PR (#782 )
• refactor: put node hasher in charge of labeling nodes and determining kinds #781
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.

[pepyakin]

splendid

[pepyakin]

Merge activity

• Feb 7, 5:25 AM EST: A user started a stack merge that includes this pull request via Graphite.
• Feb 7, 5:29 AM EST: Graphite rebased this pull request as part of a merge.
• Feb 7, 5:31 AM EST: A user merged this pull request with Graphite.