Query Processing in Encrypted Cloud Databases
Akshar Kaul · 2013
With the increasing popularity of cloud computing, a new model known as Database-asa-service (DAAS) model has emerged in recent years. In this model a service provider provides the database as a service to its clients over internet. One of the biggest problem in DAAS is “How to execute queries efficiently at the cloud resident server while maintaining data security”. In this work we present PhantomDB, which is a new framework for solving this problem. PhantomDB maintains data security by encrypting the data before storing it on the server. Each data item is encrypted under multiple encryption schemes. The Encryption schemes used by PhantomDB are carefully chosen to allow efficient query processing at the server. PhantomDB introduces the concept of Arithmetic Engine and Round Communication, which allow it to support all the standard SQL constructs with the exception of similarity operators. PhantomDB is able to support 16 out of 22 TPCH Benchmark queries. We also present a new hybrid storage model which takes unique features of PhantomDB into consideration while deciding the layout of relational data on disk. It helps in improving the performance of PhantomDB as compared to other current storage models.