Loading...

Query Correctness Assurance for Outsourced Databases

Noferesti, Morteza | 2011

590 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 42139 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Jalili, Rasoul
  7. Abstract:
  8. In the secure data outsourcing scenario, verification of the reply of an unreliable server includes assessing the authenticity, completeness and it’s integrity. In this thesis, an efficient method, with emphasis on freshness, has been introduced to evaluate the correctness of the replies from a server. It takes in hand different application needs, inherent differences in the data, and different update mechanisms. This method evaluates freshness by using timestamps alongside the data being out sourced. Due to the requirement of verifying not only the freshness of the response, but the correctness of the timestamps as well, two general methods for evaluating and verifying the responses were used, one based on digital signatures, and the other based on the Merkle Hash Tree. In the first method, the rows in the database are signed with the timestamp. While the second approach, a Merkle Hash Tree is created for each row in the database. Comparing and evaluating the differences between these two methods, a more efficient method for evaluation has been suggested. In this new method, the data is categorized in different classes based on the different time slots in which it has been updated, and a Merkle Tree has been created for each class. We’ve named this method Partitioned Merkle Tree. Using this method we are able to efficiently verify the returned response, regarding different applications requirements, nature of data, and synchronization mechanisms. Different simulations were run to show the pros and con of each of the methods in comparison which each other
  9. Keywords:
  10. Outsourcing ; Digital Signature ; Database ; Correctness ; Freshness ; Time Stamps ; Merkel Hash Tree (MHT)

 Digital Object List

 Bookmark

No TOC