Efficient optimal-parameter consistent hashing in Euclidean space

Construct a consistent hashing scheme for \(\mathbb{R}^d\) with optimal parameters that can be stored and evaluated in polynomial time in \(d\).

Background

The paper contrasts two types of consistent hashing for Euclidean space. Hash functions with near-optimal parameters are known but require exponential time to evaluate, whereas hash functions that can be stored and evaluated in polynomial time have weaker sparsity parameters. Closing this gap by obtaining both optimal parameters and efficient evaluation is explicitly identified as unresolved.