Design and Analysis of OCPA Secure Encrypted Databases A PROJECT REPORT SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Engineering

In this thesis we present the work done in the construction of new Database model eDB under the Database as a service (DaaS) framework. The main challenge in the DaaS model is to provide data security from service provider along with efficient computation of queries on the cloud database. To handle both the challenges (data security and efficient computation), we use multiple partial homomorphic encryption schemes where each scheme handles a specific database operator, and the data is encrypted under these multiple encryptions such that cipher texts of these schemes co-exist at the server. For analyzing the security of eDB, we formally define the security model (IND-OCPA) against our system and theoretically layout the arguments to prove the security of the system against the security model. We have also developed the idea of query processing for floating point numbers in encrypted database. A new efficient storage strategy in form of Encryption Dictionary is presented in this thesis. The query execution time for queries involving only order predicate clauses improved significantly in Encryption Dictionary model as to comparison with traditional horizontal storage model with a trade off on aggregate queries where it performed slightly bad.