The intrinsic intellectual merit of this proposed research is the creation of an optimized, heterogeneous optical network architecture, comprising current and future technology building blocks, that realizes the full potential of optical technology and that will be able to support exponentially increasing future bandwidth demands. Two advantages of optical networking technology over present-day networking technology that will enable the envisioned solution are: (i) the ability to offer bit-rate-, format- and protocol-independent services, which is a direct result of optical transparency; and (ii) lower capital and operational expenses in networks, owing to a significant reduction in conventional optical-electronic-optical (OEO) conversions in the network. Since optical devices behave very differently from their electronic counterparts or may not even have electronic analogs at all, it is clear that the optimum optical network architecture (in terms of cost and performance) will not be the same as the present electronic network architecture of the Internet. A key difference between current networks (and their linear extensions) and the optical networks that the principal investigators envision is the time-scale at which network reconfiguration occurs. For the foreseeable future, network reconfiguration in the optical layer will remain slow and quasi-static. However, optimally designed all-optical networks will be dynamic and require reconfiguration on much shorter time-scales. Such architectures will significantly lower the cost per bit for communication and will enable access of high rate services to the masses relatively soon, whereas electronic architectures will continue to serve only high-end users for many years to come. Developing an optimal optical network architecture will involve a restructuring and optimization of the existing network layer structure by: (i) treating architecture, protocols, and the physical layer as a single entity with strongly interacting, but distinct subsystems, and (ii) employing foreseeable technology as well as suggesting revolutionary hardware technology to exploit the benefits of optics wherever possible. The resulting intelligent optical network will be dynamically reconfigurable, and will enable various new applications by seamlessly optimizing network performance for all. Based on a system-wide optimization, the most efficient switching, routing and transport mechanisms will be developed, which will likely include electronic packet switching as an important overlay atop a much higher-speed network. The enabling architectural concepts in this research are: (i) optical flow switching (OFS) and its implications on physical and higher layer architectures, and (ii) impairment aware routing.
The broader impact of this research will be felt in broadband network applications and potentially will facilitate major advances in interactive distant learning; telemedicine; instant access to all knowledge and information (virtual libraries); and immersive virtual presence and pervasive, mobile wireless networks. Education and research will be integrated through multidisciplinary environment of this program. MIT will focus on (i) preparing a skilled and diverse workforce, including minorities and women, (ii) generating a curriculum which is research-inspired, but also industry practice-oriented, (iii) integrating engineering, technology, and business to stimulate technology transfer, and (iv) promoting education to the broader community. This program will also actively involve experts from Sycamore Networks and Cisco.
