Loading...

A Trust-based Approach for Correctness Verification of Query Results in Data Outsourcing Scenario

Ghasemi, Simin | 2012

654 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 43713 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Jalili, Rasool
  7. Abstract:
  8. One of the security issues in database outsourcing scenario is the correctness of the query results. Correctness verification includes integrity, completeness and freshness of the results. Most of the proposed approaches for correctness verification impose high overhead on the components of the scenario which prevents the scenario to implemented in practical applications. In this thesis, we have proposed a probabilistic approach which imposes acceptable overhead for correctness verification of returned results of service provider. The approach uses the previous behavior of the service provider to calculate a trust value toward it which is used to adjust the imposed overhead. In other words, this trust based approach provides a trade-off between security and performance. The trust value reflects the past behavior of service provider, as the trust value increases, the correctness verification overhead reduces and vice versa. In this thesis, we have proposed three trust based methods including trust based digital signatures, trust based fake tuples and trust based bloom filters. In trust based digital signatures method, the number of tuples form result set to be audited is adjusted according to trust value. As the trust value toward service provider increases, the number of audited tuples decreases. In trust based fake tuples method, the trust value specifies the number of outsourced fake tuples. In the trust based bloom filter method, the number of audited tuples is adjusted to trust value. In this method, we proposed a new data structure as infinite counting bloom filter which deletion of elements does not generate false negative probability. We simulated the proposed methods and compared them with some of the existing methods in correctness verification. The empirical evaluation and analysis shows that using trust concept in query result verification in addition of providing a trade-off between security and performance, reduces the imposed overheads especially the computational overheads on users and client. So in a trust based method as the trust value toward the service provider increases, the effectiveness of the method increases significantly
  9. Keywords:
  10. Correctness ; Trust ; Digital Signature ; Outsourcing ; Database ; Fake Tuples ; Bloom Filter Method

 Digital Object List

 Bookmark

No TOC