Towards Differential Query Services

Abstract

Cloud computing as an emerging technology trend is expected to reshape the advances in information technology. In a cost efficient cloud environment, a user can tolerate a certain degree of delay while retrieving information from the cloud to reduce costs. In this paper, we address two fundamental issues in such an environment: privacy and efficiency. We first review a private keyword-based file retrieval scheme that was originally proposed by Ostrovsky. Their scheme allows a user to retrieve files of interest from an untrusted server without leaking any information. The main drawback is that it will cause a heavy querying overhead incurred on the cloud, and thus goes against the original intention of cost efficiency. In this paper, we present a scheme, termed efficient information retrieval for ranked query (EIRQ) and based on an aggregation and distribution layer (ADL), to reduce querying overhead incurred on the cloud. In EIRQ, queries are classified into multiple ranks, where a higher ranked query can retrieve a higher percentage of matched files. A user can retrieve files on demand by choosing queries of different ranks. This feature is useful when there are a large number of matched files, but the user only needs a small subset of them. Under different parameter settings, extensive evaluations have been conducted on both analytical models and on a real cloud environment, in order to examine the effectiveness of our schemes.

Existing System

Existing system private keyword-based file retrieval scheme that was originally proposed by Ostrovsky. Their scheme allows a user to retrieve files of interest from an untrusted server without leaking any information. The main drawback is that it will cause a heavy querying overhead incurred on the cloud, and thus goes against the original intention of cost efficiency. Private searching was proposed by Ostrovsky et al.which allows a user to retrieve files of interest from an untrusted server without leaking any information. However, the Ostrovsky scheme has a high computational cost, since it requires the cloud to process the query on every file in a collection. Otherwise, the cloud will learn that certain files, without processing, are of no interest to the user. It will quickly become a performance bottleneck when the cloud needs to process thousands of queries over a collection of hundreds of thousands of files. As in existing work, the cloud is assumed to be honest but curious. That is, it will obey our schemes, but still wants to know some additional information about user privacy. Reference classified user privacy into search privacy and access privacy.

Proposed System

We propose a scheme, termed Efficient Information retrieval for Ranked Query (EIRQ), in which each user can choose the rank of his query to determine the percentage of matched files to be returned. The basic idea of EIRQ is to construct a privacy preserving mask matrix that allows the cloud to filter out a certain percentage of matched files before returning to the ADL. This is not a trivial work, since the cloud needs to correctly filter out files according to the rank of queries without knowing anything about user privacy. Focusing on different design goals, we provide two extensions: the first extension emphasizes simplicity by requiring the least amount of modifications from the Ostrovsky scheme, and the second extension emphasizes privacy by leaking the least amount of information to the cloud. The ADL is deployed inside the security boundary of an organization, and thus it is assumed to be trusted by all of the users. In the supplementary file available online, we will discuss how the EIRQ schemes work without such an assumption. The communication channels are assumed to be secured under existing security protocols, such as SSL, during information transfer. With these assumptions, as long as the ADL obeys our schemes, a user cannot know anything about other users’ interests, and thus the cloud is the only attacker in our security model. In our work, user queries are classified into multiple ranks, and thus a new kind of user privacy, rank privacy, also needs to be protected against the cloud. Rank privacy entails hiding the rank of each user query from the cloud, i.e., the cloud provides differential query services without knowing which level of service is chosen by the user. Rank privacy can be classified into basic level and high level, where basic level will hide the rank of each query from the cloud, and the high level will further hide the number of ranks from the cloud. Our design goal can be subdivided as follows:

Architecture