Unmanaged Internet Protocol

Abstract

Though appropriate for core Internet infrastructure, the Internet Protocol is unsuited to routing within and between emerging ad-hoc edge networks due to its dependence on hierarchical, administratively assigned addresses. Existing ad-hoc routing protocols address the management problem but do not scale to Internet-wide networks. The promise of ubiquitous network computing cannot be fulfilled until we develop an Unmanaged Internet Protocol (UIP), a scalable routing protocol that manages itself automatically. UIP must route within and between constantly changing edge networks potentially containing millions or billions of nodes, and must still function within edge networks disconnected from the main Internet, all without imposing the administrative burden of hierarchical address assignment. Such a protocol appears challenging but feasible. We propose an architecture based on self-certifying, cryptographic node identifies and a routing algorithm adapted from distributed hash tables.