Amazon.com, Inc.
Multi-party analysis of streaming data with privacy assurances

Abstract:

Systems and methods for implementing a secure and efficient cryptographic protocol for analyzing data objects while providing assurances of data privacy and security. A data object may be obfuscated and provided for analysis (e.g., to a data analytics service) without necessarily providing access to the (e.g., plaintext) data object. For example, a first computing entity and second computing entity may agree upon a function or circuit that performs a certain type of computational task, such as comparing a first data set controlled by the first computing entity and a second data set controlled by the second computing entity. An event-driven function may be invoked by the event-driven compute service in response to detecting satisfaction of a condition as part of monitoring alerts that are generated as a result of the output of the computational task described above.