OLYMPIA: A Simulation Framework for Evaluating the Concrete Scalability of Secure Aggregation Protocols

Ivoline Ngong, Nicholas Gibson, Joseph Near

Keywords: Secure aggregation protocols, Federated Learning, Privacy-preserving, Simulation

TL;DR: The paper introduces "Olympia", a framework for simulating and evaluating secure aggregation protocols in federated learning, offering insights into their real-world concrete performance and scalability.

Abstract:

Recent secure aggregation protocols enable privacy-preserving federated learning for high-dimensional models among thousands or even millions of participants. Due to the scale of these use cases, however, end-to-end empirical evaluation of these protocols is impossible. We present OLYMPIA, a framework for empirical evaluation of secure protocols via simulation. OLYMPIA provides an embedded domain-specific language for defining protocols and a simulation framework for evaluating their performance. We implement several recent secure
aggregation protocols using OLYMPIA and perform the first empirical comparison of their end-to-end running times. We release OLYMPIA as open open-source.

Submission Number: 154