This project introduces a grand challenge for the networking community and industry: reducing Internet latencies to close to the limiting physical constraint, the speed of light in free space. Achieving this vision of a 'speed-of-light Internet' could translate into significant value for today's applications; for example, shaving even a few hundred milliseconds from a Web transaction may translate into millions of dollars of revenue for a Web provider, or significantly impact emergency or defense response. The human visual system cannot correctly order visual events separated by less than about 30 millisecond; thus, a response received over the Internet within thirty milliseconds would achieve the illusion of an "instant response", potentially enabling the full potential of certain applications, e.g., tele-immersion, that have so far been limited to the laboratory. For some applications, such as massive multi-player online games, (or any online interactive community), reducing the latency of the Internet will push applications to reach a critical mass of users. A speed-of-light Internet can also be expected to spur the development of new and creative applications. This proposal aims at developing a playbook for realizing a speed-of-light Internet, by identifying the factors that contribute to latency, the applications that would benefit most, and developing an architecture, protocols and methodologies for achieving it.
The four main goals of this project are to (1) determine the impact of latency on end users, (2) design a speed-of-light Internet, (3) explore "end-point agnostic" mechanisms for reducing latency, and (4) determine how to operate, maintain, and make use of a bandwidth-constrained low-latency network. The project will investigate these questions in the context of candidate applications including electronic commerce, games, and virtual and augmented reality. The proposed work includes new algorithms and design for the wide-area network infrastructure, and new protocols and network management to deal with the dynamic partitioning of application traffic across the two infrastructures. The project will explore how to leverage low-latency networking technologies, i.e., microwave and millimeter wave radios, which have thus far been deployed as point-to-point links over long distances primarily for the financial industry; specifically, how they can be incorporated into the Internet at large. The work will be carried out in collaboration with McKay Brothers, the operators of an ultra-low-latency network linking financial markets in New York and Chicago, Akamai Technologies, a major provider of content delivery services, and Exablaze, a manufacturer of the ultra-low-latency network interface cards. This will inform the design of a cSpeed mesh network, and its interaction and interoperability with the standard Internet.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.