SAFE: Secure Aggregation with Failover and Encryption
This addresses the need for efficient and secure aggregation in federated learning, particularly for constrained platforms, though it is incremental as it builds on existing secure aggregation methods.
The paper tackles the problem of secure aggregation in cross-organizational federated learning by proposing a chain-based algorithm with encryption, resulting in a 70x performance improvement over state-of-the-art methods with 36 nodes.
We propose and experimentally evaluate a novel secure aggregation algorithm targeted at cross-organizational federated learning applications with a fixed set of participating learners. Our solution organizes learners in a chain and encrypts all traffic to reduce the controller of the aggregation to a mere message broker. We show that our algorithm scales better and is less resource demanding than existing solutions, while being easy to implement on constrained platforms. With 36 nodes our method outperforms state-of-the-art secure aggregation by 70x, and 56x with and without failover, respectively.