Authors:
Linru Zhang
1
;
Gongxian Zeng
1
;
Yuechen Chen
1
;
Nairen Cao
2
;
Siu-Ming Yiu
1
and
Zheli Liu
3
Affiliations:
1
Department of Computer Science, The University of Hong Kong, HKSAR and China
;
2
Department of Computer Science, Georgetown University and U.S.A.
;
3
College of Computer and Control Engineering, Nankay University, Tianjin and China
Keyword(s):
ORAM, Constant Communication Overhead, Oblivious Clear Algorithm.
Related
Ontology
Subjects/Areas/Topics:
Applied Cryptography
;
Cryptographic Techniques and Key Management
;
Data and Application Security and Privacy
;
Data Engineering
;
Data Protection
;
Databases and Data Security
;
Information and Systems Security
Abstract:
Oblivious RAM has been studied extensively. A recent direction is to allow the server to perform computations instead of being a storage device only. This model substantially reduces the communication between the server and the client, making constant bandwidth communication (the number of blocks transmitted) feasible. It is obvious that the larger the block size, the easier it is to construct a constant bandwidth ORAM scheme. Also, a lower bound of sub-logarithmic bandwidth was given if we do not use expensive homomorphic multiplications. The question of “whether constant bandwidth with smaller block size without homomorphic multiplications is achievable” remains open. In this paper, we show that the block can be further reduced to O(log3N) using only additive homomorphic operations. Technically, we design a non-trivial oblivious clear algorithm with very small bandwidth to improve the eviction algorithm in ORAM for which the lower bound proof does not apply. As an additional benefi
t, we are able to reduce the server storage due to the reduction in bucket size.
(More)