Secure Aggregation
Definition
Secure aggregation computes an overall result without revealing individual contributions.
A multiparty protocol that computes an agreed aggregate, often a sum of model updates, without exposing the individual contributions in plaintext to the aggregator. The guarantee applies only under the respective threat and failure model.
Simply explained
Multiple devices provide protected contributions; the collecting party sees only the combined sum under the specified conditions.
Why it matters in practice
The term contributes to secure, maintainable, and high-performance web applications as well as resilient technical decisions.
Example from business
Mobile devices provide encrypted model updates; the server sees only their sum.
Limits and pitfalls
- Equating secure aggregation with differential privacy or full anonymity, ignoring information derivable from the aggregate, or not considering collusion, participant dropouts, and metadata in the threat model.
Origin & history
Bonawitz and co-authors published a practical protocol for federated learning in 2017.
Use in mid-sized companies
- Possible example: Mobile devices provide encrypted model updates; the server sees only their sum.
Frequently asked questions
What does secure aggregation mean?
A multiparty protocol that computes an agreed aggregate, often a sum of model updates, without exposing the individual contributions in plaintext to the aggregator. The guarantee applies only under the respective threat and failure model. What mistakes should be avoided? Equating secure aggregation with differential privacy or full anonymity, ignoring information derivable from the aggregate, or not considering collusion, participant dropouts, and metadata in the threat model.
Sources
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: