Sap Se

Arne Schramm

Axel Schroepfer

Kilian Becher

Bernhard Wolf

Andre Preussner

Raik Hartung

Johannes Hoenger

Embodiments allow comparison of key figures (e.g., costs, lead times) between different entities in a privacy-preserving manner, utilizing secure multi-party computation (MPC) approaches implemented by a central service provider. The central service provider receives encrypted key figure data from each of multiple players in a peer group. In one embodiment the central service provider executes a secure computation protocol comprising a semi-homomorphic encryption scheme exhibiting an additive homomorphic property. The central service provider returns to each player, a statistical measure (e.g., top quartile, bottom quartile) allowing comparison with the other players' key figures while preserving privacy. Alternative embodiments may return to the players, a statistical measure calculated from a Boolean or arithmetic circuit implemented at the central server using other secure computation approaches (e.g., garbled circuits, secret sharing, or (semi or fully) homomorphic encryption. Embodiments may find value in maintaining privacy of key figure data shared between competitors for benchmarking.

Private benchmarking cloud service with enhanced statistics

Data streams are received for processing by a query. The query is associated with expected input data stream types and expected output data stream types. The received data streams are evaluated to determine if they correspond to the expected data stream types. If the received data stream types define data for a physical quantity that is defined in a unit of measure that is not the expected unit of measure, then unit conversion is required. Predefined mappings are utilized for performing required unit conversions and an optimal solution for conversion is determined. The unit conversions are performed automatically by unit conversion operators that are integrated within the process of generating of a runnable query. The runnable query is deployed and executed to generate result output streams that match the expected output stream type.

Automatic conversion of units of measure during data stream processing

Growing community of inventors

Dresden, Germany

Arne Schramm