Verifiable Distributed Aggregation Functions

The modern Internet is built on systems that incentivize collection of information about users. In order to minimize privacy loss, it desirable prevent these from collecting more than required for the application. promise multi-party computation data can be aggregated without revealing individual measurements collector. This work offers a provable security treatment "Verifiable Distributed Aggregation Functions (VDAFs)", class protocols being considered standardization by IETF. We propose formal framework analysis VDAFs and apply two constructions. first Prio3, one candidates standardization. VDAF based Prio system Corrigan-Gibbs Boneh (NSDI 2017). prove Prio3 achieves our goals with only minor changes draft. second construction, called Doplar, introduced this paper. Doplar round-reduced variant Poplar et al. (IEEE S&P 2021), itself candidate cost improvement modest increase in overall bandwidth computation.