An Evaluation of Software Sketches

TL;DR

This paper evaluates various Rust implementations of sketching algorithms, comparing their performance and accuracy, and finds that a hashing-based solution with Nitro sampling offers the best trade-off, providing new insights into combining sampling with filters.

Contribution

It provides a comprehensive evaluation of Rust-based sketching solutions and introduces novel insights into combining sampling techniques with Counting Cuckoo filters.

Findings

Hashing-based solution with Nitro sampling performs best

Optimal trade-off between memory, speed, and error identified

New insights on combining sampling with Counting Cuckoo filters

Abstract

This work presents a detailed evaluation of Rust (software) implementations of several popular sketching solutions, as well as recently proposed optimizations. We compare these solutions in terms of computational speed, memory consumption, and several approximation error metrics. Overall, we find a simple hashing based solution employed with the Nitro sampling technique [22] gives the best trade-off between memory, error and speed. Our findings also include some novel insights about how to best combine sampling with Counting Cuckoo filters depending on the application.