Data privacy and security have become significant concerns in today’s digital era, especially with the increasing use of cloud services. Traditionally, encrypted data must be decrypted before processing, posing a potential security risk. Apple is introducing a solution to this problem with the open-source Swift package called swift-homomorphic-encryption. Homomorphic encryption allows computations on encrypted data without revealing the underlying information.
The swift-homomorphic-encryption package allows for computation on encrypted data without the need for decrypting it or accessing the decryption key. This means clients can send encrypted data to a server for processing. The server then returns the results in an encrypted format, which the client can then decrypt. Throughout this process, data privacy and security are maintained. This technique is particularly beneficial for cloud services.
The new package leverages several advanced features. It utilizes the Hummingbird HTTP framework and offers cross-platform support thanks to Swift on Server. A Benchmark Library provides easy performance benchmarking, while Swift Crypto equips performance at a low-level cryptography primitive.
The swift-homomorphic-encryption package uses the Brakerski-Fan-Vercauteren (BFV) Homomorphic Encryption scheme based on the ring learning with errors (RLWE) hardness problem. This scheme ensures post-quantum 128-bit security, protecting encrypted data during computations and safeguarding against both classical and potential future quantum attacks.
The new package is already being utilized in Apple’s iOS 18’s Live Caller ID Lookup feature. This feature offers caller ID and spam blocking services by sending an encrypted query to a server, which then retrieves information about the phone number queried without accessing the actual phone number.
Apple’s Live Caller ID Lookup feature also employs Private Information Retrieval (PIR), which allows clients to perform private key-value database lookups. Clients can retrieve the associated value from a server using a private keyword without the server learning the keyword. This is carried out via homomorphic encryption, which requires just a minimal amount of database metadata to be synchronized with the client.
The swift-homomorphic-encryption package empowers developers and researchers to create privacy-preserving applications for the Apple ecosystem and beyond. Potential applications extend to private set intersection, secure aggregation, and machine learning. The community is encouraged to contribute to the project and to explore new uses for homomorphic encryption, all with the goal of fostering innovation and enhancing data security.
Apple’s efforts in enhancing data security and privacy through homomorphic encryption represents a significant stride in the realm of cloud services. This project can potentially revolutionize how data computations are carried out, providing greater security and preserving users’ privacy. It shows that with the right tools and methodology, it is possible to protect users’ sensitive information while fully utilizing the advantages of cloud-based services.
