Author Description

This dissertation tackles three key security and privacy challenges in database-driven DSS to pave the way for its wide development and deployment. First, the DBA relies on spectrum measurements submitted by mobile users to construct and maintain the REM, but some mobile users may be malicious or compromised to submit false spectrum measurements. To tackle this challenge, we introduce a novel mechanism for secure REM construction in the presence of false measurements. Second, crowdsourcing-based spectrum sensing relies on mobile users' participation, who not only require strong incentive, but also demand privacy protection. To tackle this challenge, we design an incentive mechanism that simultaneously achieves differential bid privacy, truthfulness, and high REM accuracy. Third, an effective approach to process a large number of spectrum access requests with low latency is to adopt the edge computing paradigm by having the DBA continuously pushes the spectrum availability updates to distributed local edge servers, which in turn process spectrum access requests from nearby SUs on the DBA's behalf. However, edge servers owned by different entities cannot be fully trusted to process SU's spectrum request based on authentic and the most recent spectrum information, which may result in either loss of revenue or harmful interference to PUs' transmissions. To tackle this challenge, we propose a novel freshness authentication mechanism to allow SUs to verify that their spectrum-access requests are decided based on authentic and up-to-date spectrum availability information.