The developers behind the Brave open source web browser have revealed a new privacy-preserving system called the Data Query and Search System. Frodo PIR.
The company says it aims to use the technology to build a wide range of use cases such as secure browsing, scanning passwords against compromised databases, certificate revocation checking, and streaming.
This scheme is called FrodoPIR. This is because “a client can make hidden queries to the server just like Frodo was hidden from Sauron.” Lord of the Ring.
PIR (short for Private Information Retrieve) is a cryptographic protocol that allows a user (aka client) to retrieve information from a database server without revealing to its owner which elements were selected.
In other words, service providers can query the platform for information (e.g. cooking videos) without inferring from a user’s search history to provide personalized recommendations or targeted advertisements based on search criteria. The goal is to make
One way to achieve this is by using an approach called homomorphic encryption. This allows computations to be performed directly on the encrypted data without requiring access to the private key.
However, the general problem plaguing such methods is that they are “expensive, either in terms of bandwidth or the time it takes to process each client query”, making practical deployment impossible. am.
This is where FrodoPIR intervenes. This includes his two phases, an offline preparation step and an online step where the client sends encrypted queries to the server.
The server then chooses to return a positive or negative value depending on whether the query was found in the database without knowing what the user was actually querying.
“In terms of database performance with 1 million KB elements, FrodoPIR takes less than 1 second to respond to client queries, with an expansion factor of over 3.6x for server response sizes, and a financial The cost is ~$1,” Brave said in the project description on GitHub.
Two Google Open Source Privacy Enhancing Technologies (PET)
The development comes as Google says it will open source two privacy-enhancing technologies (PETs) as part of its ongoing effort to democratize access to technologies beyond federated learning and differential privacy. is done on
It consists of a new machine learning tool called Magritte designed to blur objects such as license plates that appear in videos, and efficiency improvements to the Fully Homomorphic Encryption (FHE) Transpiler.
Transpilers, also known as source-to-source compilers or translators, are designed to perform computation-based queries over encrypted information without access to personally identifiable data.
PETs “offer new ways for the broader developer community (researchers, governments, nonprofits, businesses, etc.) to deploy and enhance privacy features in their own work,” Google said. I’m here.
