Research On Secure Retrieval Scheme For Structured Database Based On Homomorphic Encryption

Cloud storage has the advantage of high efficiency and convenience.The data processing capability of servers is much better than local computers.More and more users choose to store their data in servers instead of personal computers.However,data breaches are common in recent years.Data encryption is the most reliable method to solve this problem,but traditional encryption algorithms increase the difficulty of ciphertext retrieve.Meanwhile,calculations cannot be directly implemented on ciphertext.Since numeric data and character data are most frequently used in structured database,it is necessary to design a reasonable and effective ciphertext retrieval scheme.This paper proposes a secure retrieval scheme based on homomorphic encryption for the server-side structured database.With the advantage of this scheme,ciphertext in server under unreliable storage circumstance can be efficiently retrieved without exposition of plaintext.Users can also have fine-grained access to their data.The main works of this paper are described as follows:(1)Based on the analysis of data type in structured database,a double homomorphic encryption and decryption algorithm for numerical data is proposed.Meanwhile,an improved homomorphic encryption and decryption algorithm for character data is also proposed.The homomorphism of these two algorithms allows the server to operate on ciphertext directly,like calculations for numerical data and groupings for character data.With the prerequisite of sensitive data security,more calculations and matching works can be done on the server.It effectively reduces the amount of data transferred back to the client.(2)A homomorphic equivalent query and range query algorithm is proposed for the retrieve of numeric data ciphertext.Retrieval keywords are encrypted with the improved homomorphic encryption algorithm on the client.SQL statements are rewritten according to the retrieval requirements.Then the server retrieves ciphertext through SQL statements and only sends the matched results back to the client,which efficiently reduces the cost of time.In the whole process of retrieve,the server does not need to know the decryption key and it ensures the security of data.(3)Based on dual encoding function,a counting joint index generation algorithm with upper limit is proposed for fuzzy retrieve of character data.The plaintext cannot be inferred by indexes because of the irreversibility of hash function and the limited counting strategy.A two-phase character fuzzy retrieve algorithm based on joint index is designed to improve the efficiency of irrelevant data filtering on server and reduce the waste of time and space.The practicability of the proposed structured database secure retrieval scheme is proved by correctness test.In the performance test,the time efficiency of the proposed numerical equivalent query algorithm is about 16% higher than the value-based equivalent query algorithm.The numerical range search algorithm in this paper is 1.13 times faster than the method using bidirectional detection.Compared to the character fuzzy retrieval algorithm based on security enhancement,the method in this scheme increases the filtering efficiency of unrelated data on the server by 9%.