| With the increasing use of web services, many new challenges concerning data security are becoming critical. Data or applications can now be outsourced to powerful remote servers, which are able to manage outsourced source and provide service on behalf of the owners. Outsourcing data or applications to remote hosts (publishers) relieves owners from storing and managing the data or applications. Hence outsourcing reduces the overhead to provide services and helps improve service availability and scalability. However, outsourcing introduces new security and privacy concerns for the outsourced data or applications, for the hosts, as a party independent from the owners and users, may be malicious. Without techniques to protect data and applications, owners with IT needs are reluctant to outsource. This researcher proposes a one-server tree-structured data hiding scheme to protect content privacy and access privacy of outsourced tree-structured data and shows that when data is uniformly accessed, the access-redundancy and node-swapping techniques can help achieving access privacy. However, when data is not uniformly accessed, this technique is susceptible to intersection-based traffic analysis attack. This researcher then proposes access-clustering and chain-inerging techniques to reduce the risk of such traffic analysis attacks. Compared with other techniques to address access privacy, these techniques are less costly and hence more practical. Considering the importance of XML for organizing and exchanging data over the Internet and its inherent tree structure, this researcher then discusses how to apply the above techniques to build an XML data store (pXStore) for secure and privacy-preserving XML data outsourcing. Finally, this researcher discusses how to utilize these techniques for privacy-preserving outsourcing of arbitrary structured data. |
