**Session Date/Time:** 27 Mar 2023 04:00

# ppm

## Summary

This session of the Privacy Preserving Measurement (ppm) working group included presentations and discussions on the distributed aggregation protocol (DAP), the Star protocol, and secure partitioning techniques. Key topics included updates to DAP, proposed optimizations to DAP for single-helper configurations, performance measurements of Star, and an introduction to secure partitioning strategies for scalable MPC deployments.

## Key Discussion Points

*   **Distributed Aggregation Protocol (DAP) Updates:** Tim presented updates on the DAP draft, including changes to align with the latest version of the Distributed Aggregation Functions (VDAF) specification. Changes involve nonce generation, verification key negotiation, and aggregation parameter validation.
*   **DAP Single-Helper Optimization:** Chris and Brandon proposed optimizing DAP to support only one helper, reducing network round trips and latency. This sparked debate regarding the loss of generality and potential future use cases requiring multiple helpers.
*   **Star Protocol Performance:** Sophia presented performance measurements of the Star protocol, focusing on the impact of the key parameter 'k' and the choice of cryptographic curves on the scheme's efficiency. Discussed practical performance of implementations.
*   **Star Adoption:** Eric requested adoption of the Star protocol in the working group to move towards standardization. Discussion focused on identifying use cases, particularly regarding input entropy, and the properties of the privacy provided.
*   **Secure Partitioning:** Philip presented secure partitioning techniques for distributing client data across shards in MPC deployments. Strategies included F-oblivious evaluation and sparse partitioning.

## Decisions and Action Items

*   **DAP Single-Helper Optimization:**
    *   The working group will defer a decision on the single-helper optimization. It was agreed that more measurements are needed to evaluate the performance impact.
    *   Chris, Brandon, and their teams will implement the one-helper version of DAP and conduct experiments to measure performance and report back to the working group.
*   **Star Adoption:**
    * The working group will return to the list with the adoption proposal.
    * Ben will put together a document of use cases for STAR.
    * Sophia, David will formalize the differential privacy parameters of STAR.
    * The working group to review objections.

## Next Steps

*   **DAP:** Tim and other DAP editors will continue to work on the DAP draft, incorporating feedback and updates to VDAF. Future work includes scalable implementations of popular one.
*   **Star:** Continued evaluation of the implementation by Sophia, David.