Evolving the Internet Architecture Through Naming

Abstract

Challenges face the Internet Architecture in order to scale to a greater number of users while providing a suite of increasingly essential functionality, such as multi-homing, traffic engineering, mobility, localised addressing and end-to-end packet-level security. Such functions have been designed and implemented mainly in isolation and retrofitted to the original Internet architecture. The resulting engineering complexity has caused some to think of 'clean slate' designs for the long-term future. Meanwhile, we take the position that an evolutionary approach is possible for a practical and scaleable interim solution, giving much of the functionality required, being backwards compatible with the currently deployed architecture, with incremental deployment capability, and which can reduce the current routing state overhead for the core network. By enhancing the way we use naming in the Internet Architecture, it is possible to provide a harmonised approach to multi-homing, traffic engineering, mobility, localised addressing and end-to-end packet-level security, including specific improvement to the scalability of inter-domain routing, and have these functions co-exist harmoniously with reduced engineering complexity. A set of proposed enhancements to the current Internet Architecture, based on naming, are described and analysed, both in terms of architectural changes and engineering practicalities.