Agile Optical Networks

The successful execution of the current high-capacity optical-networking environment, which supports supercomputing centers at the DOE national laboratories, requires adpative and finer-grained management/set-up of necessary bandwidth and low-delay paths between distributed applications. Whereas manual management techniques have been implemented in test beds and moderate-capacity networks, the approach is not sufficient for managing large-scale networks, such as the ESnet and the National Lambda Rail, that connect tens of supercomputing centers with thousands of super servers and several thousands users. This project will develop an automated, distributed-control system for large-scale networks, along with associated tools for its users and operators. Phase I will develop a statement-of-objectives document, in which the end-to-end aspects of networking will be documented, user needs will be formulated, and measurement/monitoring needs will be defined. A strategy for managing network components (including MPLS, GMPLS, Lambda Switching, SONET cross-connect, add-drop-multiplexing, network path optimization, scheduling, and flow control) from a control plane will be defined. Commercial Applications And Other Benefits as described by the Applicant: In addition to the DOE application, the technology should be applicable to distributed large scale computing and network configurations in the DoD, NASA, universitiy research, NIH, bioinformatics, search engine warehouses, and print media