MIX-Crowds, an Anonymity Scheme for File Retrieval Systems

In this paper, we propose an anonymous scheme for file retrieval systems, MIX-Crowds, in which it is harder for an attacker to identify the requester of the file by making use of the idea of MIX [7] and Crowds [20] to establish a path from the requester to the file holder. Result shows that predecessor attack [26] is much more difficult to succeed compared with Crowds [20]. We are able to reduce the estimated number of rounds needed for successful predecessor attack for MIX-Crowds. We also propose a file transfer strategy according to file size. With such strategy, requests for small size files can be completed faster while the downloading time of large size files only increases slightly.