Journals Proceedings

Abstract

Hashing of data blocks is a common approach in peer-to-peer and Internet based distributed file systems as a way to simulate a virtual hard disk over a set of servers. We consider the case where a malicious adversary attempts to overload a server assuming that this adversary ―knows‖ how the underlying hashing scheme works. We overcome this problem by using Universal hashing a known theoretical solution. In this work we how to port such a complex hashing scheme in real implementations of Internet based file systems. In this respect our work also contributes to the well known problem of VOD (video on demand) over the Internet. The implemented system (IFS) supports cooperative caching and for concurrent read/write operations it supports Weak-Consistency. Our results suggest that Universal hashing based distributed file systems can be extremely efficient for driving Internet file systems and file systems for large clusters.