Investigation And Improvement Of Order-Preserving Encoding In Encrypted Database

Cloud computing has been the mainstream of modern internet service deployment because of its low cost and convenience.But it is insecure that cloud holds many personal sensitive information and privacy.So there is an urgent need to research how data privacy could be protected especially in untrustworthy cloud.Encrypting records in database can protect privacy.But operations on encrypted data cannot be efficiently executed,such as range queries and sort operations.It is hard to balance security and availability.To resolve this problem,Order-Preserving Encryption(OPE)scheme were suggested,which preserves the order of plaintext.MOPE,suggested in 2013,is a mutable order-preserving encoding scheme based on binary search tree which works on all data types with ideal security.The order-preserving codes may change,and the high frequency of codes mutation will bring additional computation overhead to the server.The thesis analyzes the performance of mOPE,and then improve the process of codes mutation.A new method called cmOPE is suggested which reduces the frequency of codes mutation by means of using reconstucting complete binary search tree encoding algorithm in the process of codes mutation.For the scenario of concurrence,we designed an method called xmOPE which could encrypt and insert any number of records just in one operation.We also designed and implemented an OPE system based on xmOPE.For security,both cmOPE and xmOPE conform to IND-OCPA standard so that they only leak the order of plaintext.Theoretical analysis and empirical data show that cmOPE reduces the workload of server.Compared with mOPE,cmOPE increases the efficiency 1.5 times under the circumstance of low network latency,and it has a very low frequency of mutation.The xmOPE system has high performance on batch processing insertion requests which could adjust the amount of batch-processing records depending on the number of requests,and network latency affects it less.