Distributed Aggregation Protocol

The Distributed Aggregation Protocol (DAP) is an IETF standard for Privacy Preserving Measurement (PPM): it computes aggregate statistics without exposing individual client data.

DAP is the Distributed Aggregation Protocol, a multi-party system for Privacy Preserving Measurement (PPM). The core mechanism involves a large set of clients splitting their sensitive measurement data (e.g., user behavior counts) into encrypted shares. These shares are distributed across a pair of non-colluding servers: the Leader and the Helper (Aggregators). This architecture prevents either server from reconstructing any single user's input. The protocol leverages Verifiable Distributed Aggregation Functions (VDAFs), such as Prio3, to jointly compute the final aggregate result. Major tech players like Google and Apple have implemented early versions for use cases like exposure notifications; Cloudflare's Daphne is a notable open-source aggregator implementation.

https://datatracker.ietf.org/doc/draft-ietf-ppm-dap/