The goal of the proposed work is to build a foundation for enabling a rich set of applications on entirely unstructured, decentralized, and distributed networks. Recently, a number of different research applications have been built on similar platforms in the context of peer-to-peer systems. While these research systems offer great promise, we believe significant work is required in order to build deployable applications over these systems.
This work is motivated by the open issues in this area, and is organized into three major thrusts: (1) placement and lookup algorithms that are robust against a range of adversarial failure models, and yet - in contrast to, e.g., DHTs - do not require or assume any structure on the underlying network topology; (2) efficient and resilient techniques for sophisticated querying of P2P namespaces in unstructured environments; and (3) methods that allow peers to explicitly incorporate individual trust and privacy policies into protocol actions. The work is terms of four applications: (1) an ad-hoc public key infrastructure; (2) a robust search and query application for distributed data; (3) a censorship-resistant publishing application; and (4) an application for efficient ad-hoc access to scientific datasets. Individually, each of these applications are important building blocks that exercise different aspects of our proposed work, and directly advance the state of the art in the sciences, engineering, and national security. Together, they form a compelling set of motivating examples for the research to be conducted under this project.
