At the end of this decade, national and international-scale scientific collaborative applications will need intelligent signaling and dynamic control of very-high-performance optical networks. The software proposed here will allow scientific applications to directly control an advanced, all-optical, IP-over-wavelength metropolitan-scale network, based on Dense Wave Division Multiplexing (DWDM) and photonic switching. Optical networking technology is rapidly migrating from ultra-expensive long-haul implementations to regional- and metro-area networks. The trend is to provide a general infrastructure with a wide range of common services. However, the flexibility inherent in these new technologies provides the research community with an opportunity to move beyond general requirements and support large-scale e-science applications that require advanced networking capabilities. We propose a software development effort that will lead to 21st-century applications over 1000-fiber, 1000-wavelength photonic networks. These evolving, extreme applications requiring optical networks include high-energy physics, astrophysics, climate modeling, oceanographic modeling, architectural design, molecular modeling, industrial design, advanced photon source experimentation, materials science, and industrial engineering. Underlying such applications are cross-cutting support technologies, such as advanced digital video, remote access to scientific instruments, specialized visualization displays, data-mining, cluster supercomputing and high-performance distributed computational systems. To enable the full potential of such applications, it is not sufficient to simply provide high-performance networks; these applications need intelligent, dynamic controls to adjust network resources. The proposed software development efforts will leverage the significant potential of a newly installed metro optical testbed for application-level dynamic control of resource discovery, allocation and adjustment. Efforts at many levels are required to make such flexibility available in service provisioning, infrastructure and service resource management: Research into the behavior of advanced scientific applications, not just on extremely high-performance optical network, but on one that can be dynamically adjusted at a granular level Identify application-level networking requirements, investigate management techniques for optical networks, and study new service provisioning models related to application needs Research new methods for application signaling Investigate interconnections between application signaling and IP-based control-plane methods, such as through GMPLS Test deployment of those techniques on an advanced testbed and analyze results Experiment with multiple-service provisioning to ensure gateways to traditional networks and protocols Develop a system for performance metrics, monitoring and analysis Create a testbed for StarLight, the next-generation, optically based STAR TAP, and for other advanced research networks. The testbed for this project is an a four-node optical network, OMNInet, initially linking a core node on Northwestern University's Chicago campus with a node at the University of Illinois at Chicago, the Canadian Network for the Advancement of Research, Industry and Education (CANARIE) CA*net4 node at its Chicago Point of Presence and a node at Northwestern's Evanston campus. The sites are separated by distances of 5 to 20 miles, connected by dedicated technology trial-fiber service provided by SBC/Ameritech. Each node includes a Nortel Networks WDM photonic switch, an Optical Fiber Amplifier (OFA) and high performance router/switches. These sites will also have access to Nortel and SBC/Ameritech testing personnel, expertise, and equipment. Participants in this project, led by the Electronic Visualization Laboratory (EVL) at the University of Illinois at Chicago, include the International Center for Advanced Internet Research at Northwestern University, CANARIE, Argonne National Laboratory, MREN (Metropolitan Area Research and Education Network), Nortel and Ameritech.