Research On Untraceable Oblivious Random Access Memory Protocol For Cloud Computing

Cloud storage is an important service extended and developed in the concept of cloud computing,which has practical application value and practical significance.Cloud storage not only effectively reduces the local storage space of users,but also provides users with sufficient storage resources.However,on the one hand,when stored shared data are subject to frequent access operations,the server’s address sequence or access mode is hidden.Then,the access mode will be exposure and the privacy data will be leak.Meanwhile,the existed structures of oblivious random access memory(ORAM)have many problems,such as complex eviction operation,low utilization of storage space,and high computational cost.On the other hand,most of the existed random storage protocols adopt one-to-many online data interaction mode,users may not be able to access the required data timely.Once a user is revoked,the stored data in sever needs to be completely updated.What’s more,the server may be attacked by malicious or collusive users.Therefore,in this paper we proposed the corresponding solutions to the mentioned problems,which are shown as follows:1)Oblivious random access memory supporting dynamic operationIn this protocol,a new dynamic storage structure composed of one-way circular linked table in a binary tree and special pointer group is proposed.This structure designs a mapping relationship between the data header pointer and its corresponding address sequence.Based on the structure,a series of algorithms of access operation and eviction operation are designed,which makes the dynamic operation of data faster and more efficient.Meantime,we also design a ORAM protocol based on the dynamic storage structure,which optimizes the average bandwidth and the worst block of the server,effectively hides the access mode of the data,and prevents the malicious server from tracking the address sequence.From the results of theoretical analysis and experimental comparison,it can be seen that the protocol provides a security guarantee for users inadvertently accessing data from the server.2)Oblivious random access memory supporting privacy preservingIn this protocol,we propose an ORAM protocol that supports privacy protection and adopts a many-to-many offline data interaction mode.First of all,an identity based proxy reencryption algorithm is proposed between the agent and the user to realize the offline query of data and protect the user’s privacy data.Secondly,the protocol supports users’ lazy update.Once users are revoked,only the data stored in the server needs to be updated,which greatly saves the computing cost of the server.The results of numerical analysis and experimental comparison show that the proposed protocol has low computation and communication costs,high utilization of storage space and low computation costs on the client and server side,which can ensure users to access data offline and meet the needs of group members for privacy data protection.