Network Virtualization Mechanisms for Mobile Communication

Columbia University, New York NY

Investigators

Abstract

This project focuses on creation and deployment of the new network virtualization mechanisms to support mobile communication and, in particular, efficient transparent migration of end-to-end live network connections without any changes to existing network protocols. To this aim, novel mobile communication architecture is proposed, called Virtual Network Address Translation (VNAT), which enables connection mobility for a spectrum of computation units, ranging from a single process to the entire host. Three key mechanisms are used to enable transparent live connection mobility: connection virtualization, connection translation, and connection migration. Connection virtualization virtualizes end-to-end transport connection identification by using virtual endpoints rather than physical endpoints (e.g., IP addresses and port numbers). As a result, connection identifications no longer depend on lower layer network endpoints and are no longer affected by the movement of network endpoints. Connection translation translates virtualized connection identifications into physical connection identifications to be carried on the physical network. As connections migrate across the network, their virtual identifications never change. Instead, they are mapped into appropriate physical identifications according to the endpoints' attachment to the physical network. Finally, VNAT connection migration keeps states and uses protocols to automate tasks for connection migration such as keeping connection alive, establishing a security key, locating migrated end-point (s), and updating virtual-physical endpoints mappings. In this project, the VNAT architecture is implemented and is fully compatible with, and does not require any modifications to, existing networking protocols, operating systems, or applications. The results of this research will have a fundamental scientific impact by providing crucial networking mechanisms for future innovations in mobile networking research.