The advancement of network and storage technologies has generated rapid growth in the field of distributed storage over the past few years. Strong demands for more and improved storage has developed from a growing economic and cultural need to store and archive as much data as possible.
The current client/server architectures of most storage systems are not suitable for large-scale global storage solutions. They have a major drawback: as the number of clients increase, the load and bandwidth demands on the server also increase, eventually preventing the server from handling additional clients. The P2P information-sharing model has significant potential in the design of large-scale repositories due to its inherent scalability, decentralized architecture and low cost of entry [10]. The recent proliferation of P2P technologies, such as Napster, Gnutella, Freenet and Free Haven provide working examples of the power of the P2P model to facilitate global information sharing.
This proposal addresses this need for a secure, scalable storage system. We evaluate current peer-to-peer (P2P) technologies for the development of a platform to provide a secure, distributed object repository for an enterprise environment. We describe the characteristics that are needed to make such a storage system feasible and explain why the P2P architecture needs to be considered. We then determine which of these attributes are met by some of the most successful P2P implementations today and point out the characteristics that are lacking. Finally, we introduce ideas for a unique, complete, secure, P2P object repository that borrows from many of the strengths of some of the current popular P2P implementations.
Benefits: We intend to apply P2P secure solutions to Internet Storage in general, and in particular focusing on Internet Backup. We will work with Storage Networks (SSP) and Connected, an Internet Backup Company for Compaq Computers