| Nowadays, with the rapid growth of the Internet and the bloom of the e-commerce activities, all kinds of network facilities are quickly progress. Database, as the storage systems of information, are one of the important cornerstones of IT infrastructures and provide the data storage and sharing services. Database is an indispensable access resource for retrieving the up to date information. However, the retrieval content of user from database or web services is sensitive, such as patent databases, medical databases, and stock quotes. These inquired contents may take the user’s private information or interest exposure to the provider of database at any time and place. Therefore, it’s vita for protecting the user’s private information when retrieve records form database. Private Information Retrieval (PIR) is that allows a user to retrieve one of records from database while hiding the identity of the records form the database server, so as to prevent their retrieval private information being leaked. Private Information Retrieval was put forward concept and model for the first time in1995by Chor, and given a standard definition in1998:a database is modeled an w-bits strings d1d2...dn,i∈{1,2...,n}is the data index. A user sends to server a query q{i) to retrieve the i-th data item di, the database server sends the di to user without knowing any information about’i’. Since then the problem of Private Information Retrieval has attracted a considerable amout of researcher attention. An efficiency of PIR protocols is typically measured by their communication complexity and computation overhead necessary to answer a query. Private Information Retrieval is not only to protect the user’s retrieveal privacy, but also can be used for public key encryption system, anonymous authentication, PIR-Wrinting and all kinds of other situations.Protecting the user’s retrieve privacy is very important. However, with the popular of the cloud storage service, more and more users would like to store their encrypted information on cloud darabase. When the user needs to update the records in database, it’s essential to prevent the update privacy of users revealing to sever. PIR-Wring is that user updates one element of the database without revealing to the database provider which element is updated and to which value. PIR-Writing is one important branch of Private Information Retrieval. The first effectively PIR-Writing protocol was proposed in2007by Boneth, D. The data model of PIR-Writing is that private data f={f1,f2,...fn} was encrypted (by users) and stored in the database, andi∈{1,2...,n}is the data element index. Each element is l-bits. User updates the i-th element fiof the database to y, while the database server has know nothing about the information‘i’, also nothing to know the uodate contenty.Because of the in-depth study gradually of Private Information Retrieval, according to the security models, we group existing PIR shemes roughly into three catagories:Information Theoretical Private Information Retrieval,(IPIR); Computational Private Information Retrieval,(CPIR); Hardware-based PIR,(HW-PIR). But, most current PIR protocols are ill-suited to practical or large database due to the higher communication and computation complexity (relate to the size of database n). In the semi-honest model, and from the practical and security view, the main work of this paper is as follows:1. Based on the secure coprocessor, and combined with the security of the difficult calculation in CPIR, we propose a new HW-PIR method which used for searching in encrypted database. Through bit stream permutation and proxy re-cryptography ideal, it keeps the almost high efficiency O(1) in online inquires stage, and reduces the computational cost of secure coprocessor. This is a strong secure method that not only ensure the user’s privacy information not be leaked to database server, but also prevent inquires information of user and database privacy being revealed by SC and malicious attacks. Specially, SC can load other balance host so that it is can reduce the database burden. This model has a great practical value for the database which needs to be frequently inquired.2. We futher to study the PIR-Writing. The current PIR-Writing protocol still has the higher communication complexity (O(klogn+kl)) and computation complexity (O(nlogn+ln)), and most of them are appropriate for signal user’s database scenario, such as KuaiPan, Hotmail service. We proposed a new PIR-Writing method based on HW-PIR model that keep the higher communication complexity O(logn) and improve the online updating computation cost of PIR-Writing from O(nlogn+ln)to O(l), thus it can decrease the update repond time of database and have the practical value. In particular, our method is not only appropriate for signal user’s database scenario, but also support the application that the database information is shared by multiple users. |
PDF Full Text Request