An association probability based noise generation strategy for privacy protection in cloud computing

Publication Type:
Conference Proceeding
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2012, 7636 LNCS pp. 639 - 647
Issue Date:
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail2012001284OK.pdf370.42 kB
Adobe PDF
Cloud computing allows customers to utilise IT services in a pay-asyou- go fashion to save huge cost on IT infrastructure. In open cloud, 'malicious' service providers could record service data from a cloud customer and collectively deduce the customer's privacy without the customer's permission. Accordingly, customers need to take certain actions to protect their privacy automatically at client sides, such as noise obfuscation. For instance, it can generate and inject noise service requests into real ones so that service providers are hard to distinguish which ones are real. Existing noise obfuscations focus on concealing occurrence probabilities of service requests. But in reality, association probabilities of service requests can also reveal customer privacy. So, we present a novel association probability based noise generation strategy by concealing these association probabilities. The simulation comparison demonstrates that this strategy can improve the effectiveness of privacy protection significantly from the perspective of association probability. © Springer-Verlag Berlin Heidelberg 2012.
Please use this identifier to cite or link to this item: