In an era where data is the new gold, ensuring its privacy and security has never been more critical. Secure computation, is a powerful branch of cryptography, allowing companies to perform computations on sensitive data without revealing the actual information being processed. In this blog, we’ll explore what secure computation is and how it’s used to protect consumer data.
What is Secure Computation?
Secure computation is a cryptographic technique that enables multiple parties to jointly compute a function over their individual inputs while keeping those inputs private. This is known as "encryption in use" because the underlying data remains encrypted while it is being processed on remote servers or in the cloud.
The primary goal of secure computation is to ensure the confidentiality, integrity, and privacy of data throughout the computation process. It accomplishes this without relying on a trusted third party, making it particularly valuable in scenarios where data sharing and privacy are paramount. This means that two or more parties can collaborate on data analysis or computations without exposing their sensitive data to one another.
How are Media Companies and Brands Using Secure Computation to Collaborate?
Secure computation is applied in a range of scenarios where privacy and data security are paramount. Naturally, secure computation is a great fit for data sharing and collaboration among publishers and advertisers.
Both publishers and advertisers can benefit from a type of secure computation called Private Set Intersection (PSI) protocol. It allows two or more parties to compute the intersection of their private datasets without revealing any information about the records not in the intersection. Optable, for instance, provides an open-source matching utility that allows partners of Optable customers to securely match their first-party data sets with them using a PSI protocol.
How does secure computation work?
Secure computation can be implemented in two main ways: 1) via pure cryptography (using Fully Homomorphic Encryption (FHE) and Secure Multi-Party Computation (MPC)) or 2) through secure hardware (using Trusted Execution Environments (TEEs).
Fully Homomorphic Encryption
FHE is an incredibly powerful tool for protecting data privacy in the digital age. It enables analytics to be performed on encrypted data without ever having to decrypt it. The ad tech industry can certainly benefit from full-scale analytics without the risk of exposing personally identifiable information (PII).
While FHE has the potential to revolutionize the advertising ecosystem, it is unfortunately quite computationally intensive and limited in its current capabilities. Therefore it is not yet ready for widespread adoption. There is ongoing research to make FHE more efficient and functional in the future.
Secure Multi-Party Computation
MPC is a form of secure computation that uses a cryptographic protocol to enable two or more businesses with private data to perform a joint computation while keeping their individual inputs private. Each entity only learns what can be inferred from the computation result.
Often, the secure computation part is outsourced to two helper servers. Before data leaves a user's device, it is encrypted to both helper servers, which decrypt it partially and perform computation on the partially encrypted data. Neither server is ever able to see the original user data.
MPC protocols provide a high level of security but come with a tradeoff. They require sophisticated cryptographic operations which incur higher computation and communication costs. This makes this technology tailored for specific tasks, which can get very expensive.
How Does Optable Use MPC?
In the past year, Optable has been a leading contributor to the IAB Tech Lab’s Open Private Join and Activation (OPJA) that enables interoperable privacy safe ad activation based on PII data. At the heart of OPJA is a secure match using a PSI protocol that allows advertisers and publishers to match their PII data. One of the ways to perform this match is using MPC — the respective clean room vendors act as the MPC helper servers, which jointly compute the overlap without ever learning the identifiers not in the overlap.
In an age where data privacy is a growing concern, secure computation emerges as a vital technology that plays an important role helping companies comply with data protection regulations while still fostering innovation and cooperation among business partners.