Probing Distribution in Time and Space for IP Alias Resolution

Abstract

The Internet is composed of thousands of networks, interconnected to provide end-to-end IP (Internet Protocol) connectivity. However, Very little public information is provided about these networks and their interconnections. The information needed to create an Internet map of the routers and the links between those routers must be derived from techniques for discovering IP addresses (traceroute) and for associating IP addresses that belong to the same router (IP aliases). Both processes IP address discovery and IP alias resolution require a large measurement infrastructure, and they introduce variable amounts of traffic into the network. Although systematic proposals have been made for creating a scalable IP address discovery system, the equivalent system for resolving IP aliases is far from being determined. In this paper, new proposals for obtaining a scalable IP aliasing system are evaluated and compared with existing solutions. Distributing measurements along multiple vantage points (spatial distribution) and extending probing tasks over time (temporal distribution) have been identified as the key methods for reducing the overhead of IP alias resolution.