What is Confidential Computing

Confidential Computing unveils data insights by securely opening data silos. Using Multi-party Computation (MPC) and secret sharing the data can be analyzed while it remains encrypted.

Unlike traditional methods that require a trusted party with full data access, Confidential Computing allows joint analysis where each computing party only holds an encrypted portion of the data. Using secret sharing, a technique that encrypts data by splitting it into multiple parts, parties can compute on the encrypted data while the original data cannot be reconstructed by anyone - not even by those performing the analysis. This ensures security as all shares are required to reconstruct the original data, guaranteeing privacy as long as one participant is trusted.

An example: Combining health and location data

Suppose a health official thinks that easy access to public transport lowers the incidence of cardiovascular disease. To test this idea, health data needs to be combined with residential data. However, data privacy regulations make this type of analysis very challenging to conduct.

Traditionally, data from different sources, such as the health ministry and the statistics agency, are held in data silos due to privacy concerns. This makes it difficult to perform joint analysis.

Confidential Computing Solution

Confidential computing offers a solution to this problem. Using MPC, data can be analyzed while remaining encrypted. This would allow the health ministry and the statistics agency to collaborate without compromising data privacy.

Without Confidential Computing, it can take months to go through the process of filling out forms and getting approval to use the unencrypted data. Confidential computing significantly speeds up the process thanks to its inherent secure protocol, which uses only encrypted data. This process potentially reduces approval times from months to hours or days, a major improvement for time-sensitive research.

How it Works: Secure Data Sharing and Analysis

1. Within the data providing process, each agency would install a secure data runway within their own environment. The runway encrypts data using secret sharing, splitting it into encrypted fragments and distributing them across multiple computing parties, each of which runs and MPC node.

2. These parties execute the confidential computations, using blockchain technology to orchestrate the MPC protocols. This allows the MPC nodes to provide accurate results while the original data is never revealed during the process.

3. The computation results are compiled into a report and delivered to designated recipients.