Thumbnail Image

Privacy-preserving encoding for cloud computing

Information in the cloud is under constant attack from cyber criminals as profitability increases; user privacy is also at risk with data being mined for monetary value – the new gold. A single leak could have devastating consequences for a person or organisation, yet users have limited control over their privacy. It is becoming clear that the current model for public cloud computing is flawed, where cloud vendors and their employees can no longer be trusted to protect user data. Privacy-preserving computation in the cloud keeps data private at all times but still remains functional, thus returning control of data back to users. The cloud could then perform operations using data that it cannot comprehend. The end-user would then be able to retrieve the results from the cloud and unlock the real answers. Homomorphic encryption is a solution for privacy-preserving processing, allowing computation over cipher text. At the time of writing, a fully homomorphic system allows arbitrary operations but requires minutes to compute an operation, whereas partially homomorphic encryption can only support a single operation, meaning it cannot be a generic solution to privacy-preserving computing. Another solution is multi-party computation, which uses a distributed approach built upon homomorphic encryption but currently suffers other limitations like reusability and lacks the ability to be truly dynamic. The primary objective of this research is to design a solution for the cloud that offers privacy-preserving data computation but provides performance and flexibility. A novel approach for multi-party computation is developed, where the combination of encoding and distribution helps provide the balance between security, performance and utility. Privacy is maintained by each distributed entity only receiving a small portion of the actual data through encoding, where attempting to brute-force the data results in a vast number of possibilities, similar to encryption. Functions are defined with universal or custom logic and are computed quickly, as the performance overhead is no longer computational but network latency. A cloud voting application was used for analysis between existing solutions and the novel approach taken by this research, which is able to add thousands of votes per minute, giving practical privacy-preserving processing in the cloud.
Type of thesis
Will, M. A. (2019). Privacy-preserving encoding for cloud computing (Thesis, Doctor of Philosophy (PhD)). The University of Waikato, Hamilton, New Zealand. Retrieved from https://hdl.handle.net/10289/12384
The University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.