Despite its enormous success, today's Internet suffers certain well-known shortcomings, and is increasingly strained to meet the high availability, reliability, mobility, manageability and security demands of modern applications and services. In contrast, Ethernet is largely plug-&-play; however, this traditional layer-2 technology can hardly meet the scale as well as the efficiency and robustness requirements of emerging and future large, dynamic networks such as data centers and cloud-computing services.
In this project, the PI is developing VIRO --- a novel and paradigm-shifting approach to network routing and forwarding. VIRO simultaneously addresses the challenges faced by IP networks and Ethernet by decoupling routing from addressing and by integrating and unifying routing and forwarding performed by the traditional layer-2 and layer 3. VIRO introduces a topology-aware, structured virtual identifier (vid) space onto which both physical identifiers and higher layer addresses/names are mapped. It employs innovative DHT-style routing mechanisms to build end-to-end connectivity from bottom-up, and to route/forward data packets using vids only. Hence, VIRO is not only highly scalable and robust, but also namespace-independent.
This research project has the potential to reduce the complexity and costs of operating and managing future large, dynamic networks, and to foster creation and deployment of new Internet services. The PI plans to disseminate the research outcomes through publications and outreach activities. VIRO prototypes will be implemented and made publicly available. Through participation in future Internet design forums and collaboration with industrial partners, the PI will seek to impact the larger industry behind Internet technologies.
