Minghao Zhao

Forward Private Searchable Symmetric Encryption with Optimized I/O Efficiency

Recently, several practical attacks raised serious concerns over the security of searchable encryption. The attacks have brought emphasis on forward privacy, which is the key concept behind solutions to the adaptive leakage-exploiting attacks, and will very likely to become a must-have property of all new searchable encryption schemes. For a long time, forward privacy implies inefficiency and thus most existing searchable encryption schemes do not support it. Very recently, Bost (CCS 2016) showed that forward privacy can be obtained without inducing a large communication overhead. However, Bosts scheme is constructed with a relatively inefficient public key cryptographic primitive, and has poor I/O performance. Both of the deficiencies significantly hinder the practical efficiency of the scheme, and prevent it from scaling to large data settings. To address the problems, we first present FAST, which achieves forward privacy and the same communication efficiency as Bosts scheme, but uses only symmetric cryptographic primitives. We then present FASTIO, which retains all good properties of FAST, and further improves I/O efficiency. We implemented the two schemes and compared their performance with Bosts scheme. The experiment results show that both our schemes are highly efficient.

Multi-user searchable encryption with a designated server

Multi-user searchable encryption (MSE) enables authorized users to search over encrypted documents in the cloud. Generally, security problems in existing MSE schemes are solved as follows: (1) transmitting authority values and search tokens through secure channels to resist keyword guessing attack; (2) involving a trusted third party (TTP) to manage users and (3) relying on online users to distribute the decryption keys. However, these methods result in extra overhead and heavily restrict the scalability of the systems. In this paper, we propose a secure channel-free and TTP-free MSE scheme. It is secure against keyword guessing attack by introducing a designated server. And it achieves fine-grained access control to grant and revoke the privileges of users without TTP. More specifically, each document is encrypted with a unique and independent key, where the key distribution is integrated with user authorization and search procedures. We provide a concrete construction of the scheme and give formal proofs of its security in the random oracle model.

An ORAM-based privacy preserving data sharing scheme for cloud storage

Abstract Data sharing is one of the basic applications for cloud storage, which is inherently suitable for scalability and multitenancy feature of cloud computing. Generally, for security and privacy concerns, clients tend to conceal (e.g. encrypt) their data content. However, access patterns, usually generated by behavior of users in sharing data rather than data content itself, may cause severe sensitive information leakage . Recently, oblivious random access memory (ORAM) has drawn increasingly attention as it is an ideal cryptographic tool for access pattern hiding. However, the existing ORAM-based data sharing schemes involve various deficiencies, either in high complexity for computation or heavy reliance of complex cryptography primitives. Inspired by the former schemes, in this paper we propose a novel ORAM based data sharing scheme with high security guarantee and high efficiency. The scheme can prevent the data block from arbitrary modification through Shuffle Correctness Proof. The security of the scheme is based on the IND-CPA security of encryption scheme, the unforgeability of Identity-Based signature and the security properties of basic Path-ORAM. Analysis shows that the scheme has the optimal computation and communication complexity.

Efficient and secure outsourced approximate pattern matching protocol

Pattern matching is a basic algorithmic problem that identifies the appearance as well as the location of a pattern in a specific text, and one of the most important variants of that, approximate pattern matching, can be used to discern a substring in the text that is similar to the pattern, as long as their differences stay within a certain threshold. It serves as a basic component in many real-world applications, such as facial recognition, DNA matching and music retrieval. Motivated by the newly emerging secure outsourced computing, in this paper we proposed protocols to realize these functionalities in a privacy-preserving manner. Specifically, we constructed exact and approximate matching protocols, and both of them ensure that the party holds the text (with length of n) learns noting about the pattern (with length of m). We composed a novel idea to combine secret sharing scheme with oblivious transfer (OT), such as to transform the secure pattern matching problem into reconstructing of a shared secret, which means that if a shared secret can be correctly reconstructed, it indicates the pattern indeed exists in the text. Our protocol for approximate pattern matching is generated in the cloud-assisted setting, where the reconstruction phase is outsourced to an honest-but-curious cloud server. Using oblivious transfer extension technique, a powerful method to use few integrated OTs to implement large-scale single OTs, our protocol is efficiently constructed. Both of the protocols are secure in semi-honest model, and we present a detailed secure simulation-based proof in this paper.

Keyword guessing on multi-user searchable encryption

Multi-user searchable encryption enables the client to perform keyword search over encrypted data while supporting authorisation management. Most of these schemes are constructed using public key encryption. However, public key encryption with keyword search is vulnerable to keyword guessing attack. Consequently, a secure channel is necessarily involved for secret information transformation, which leads to extra severe burden. This vulnerability is recognised in traditional searchable encryption, but it is still undecided whether it also exists in multi-user setting. In this paper, we firstly point out that keyword guessing attack is also a problem in multi-user searchable encryption without the supposed secure channel. By an in-depth investigation of some schemes proposed recently and simulating the keyword guessing attack on them, we present that none of these schemes can resist this attack. We make a comprehensive security definition and propose some open problems.

Longest Common Sub-sequence Computation and Retrieve for Encrypted Character Strings

Longest Common Sub-sequence is a basic algorithm problem. It serves as a basic component for a variety of applications in information processing and bioinformatics. It is a NP-hard problem and often manipulated using dynamic programming, which is relatively fast but involves large memory space. Fortunately, cloud computing and outsourced computing provides a practical method for overload alleviation. However, for the security and privacy concern, clients hope to encrypt their data before upload them to the cloud, meanwhile maintain the ability for the cloud to process on the data. In this paper, we propose a method to computing Longest Common Sub-sequence using somewhat homomorphic encryption. Beyond that, we show how to use our achievement into searchable encryption to achieve rich expressiveness.

Data Sharing with Fine-Grained Access Control for Multi-tenancy Cloud Storage System

Data sharing is one of the most significant applications of cloud computing. For security and privacy concerns, clients generally encrypt their data before upload them to the cloud. The existing data sharing schemes either entirely rely on the cloud to enforce access control or inevitably involve a trusted third party (TTP) to perform secret key distribution. This thesis proposes a secure data sharing scheme without TTP involved. Our scheme allows users to classify their data and achieves a fine-gained access authorization. The key-distribution is integrated with the user authorization and data sharing procedure. In terms of security, except for semi-honest cloud service provider and external adversary, we also take internal adversary into consideration and analysis security in this strong model.

Efficient ORAM Based on Binary Tree without Data Overflow and Evictions

ORAM is a useful primitive that allows a client to hide its data access pattern and ORAM technique as a wide range of applications nowadays. In this paper, we propose a verified version of binary-tree-based ORAM with less data access overhead. We provide a new method to reselect the leaf node and write data back to the tree, and accordingly, avoid complicated evict operation. Besides, the bucket capacity is reduced to a constant level. Overall, our scheme improves the efficiency meanwhile maintains security requirement of ORAM.