Verifiable encrypted computations
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
The ubiquitous usage of cloud computing has revolutionized the way individuals and organizations access and process their data. By delivering numerous services on the internet, cloud computing allows users to store vast amounts of data and perform heavy computations on them without the need to manage complex local computing resources. This innovative technology offers scalable, cost-efficient, and flexible access to storage, computing resources, and applications. Moreover, cloud computing allows businesses to deliver their products and services easily and effortlessly across the globe. ☐ Given that widespread usage of cloud computing and the instrumental development to the world, cloud servers are an alluring target for cyber-attackers who seek to eavesdrop on or steal users’ sensitive data. The providers of cloud computing are also another threat, where malicious cloud providers can access users’ confidential information for their own benefit, such as targeting users with advertisements. ☐ One solution to these concerns is using cryptography. Cryptography is a way to protect our data by transforming them into unreadable form that prevents unauthorized entities from accessing our data. It works well when users want to protect their data when stored on remote servers or during communication. However, this traditional cryptography falls short when we need to perform computations on the data. In this case, a user will have to reveal his encrypted data to the cloud provider to allow performing the computations on them. ☐ To address the shortcomings of traditional cryptography. People tend to use another form of cryptography called Fully Homomorphic Encryption (FHE). FHE allows performing computations on encrypted data, without the need to reveal them. In this thesis, we propose frameworks that enable non-experienced users to seamlessly use homomorphic encryption techniques to securely outsource computations to untrusted cloud providers. Furthermore, we complement this by adding the integrity component, Zero-Knowledge Protocols, so that users can verify the correctness of the computations performed, thus protecting against cheating providers. Moreover, this thesis investigates the design of new optimized functional units to allow users to perform non-linear operations on encrypted data.
Description
Keywords
Cloud computing, Cryptography, Cybersecurity, Homomorphic encryption, Zero-knowledge proofs