Ultra High Speed Transport Control Protocol and Architecture (Revised)

79438S Emerging distributed science applications are expected to generate petabytes of data that must be transferred to geographically distributed tera-scale computing facilities. Ultra-high bandwidth will be needed to achieve this transfer. While the bandwidth is available in network link technology, transport architectural inefficiencies severely limit the throughput attainable over such links. This project will develop a transport protocol that supports emerging network-host architectures and deals effectively with transport inefficiencies in ultra high speeds is required. The approach will be based on Transmission Control Protocol-Fiber, a network independent transport mechanism that allows networking at full link speeds over ultra high speed networks. It effectively deals with transport inefficiencies over such speeds, supports zero-copy and parallel network-host interface architectures, and is based on fast and accurate network measurement techniques. Phase I will provide a feasibility study of the proposed Transmission Control Protocol-Fiber technology and address the requirements of scientific distributed applications at ultra high speeds. A laboratory demonstration will show the potential and limitations of the proposed approach. Commercial Applications and Other Benefits as described by the awardee: Transmission Control Protocol-Fiber should enable the effective conduct of distributed scientific applications such as environmental monitoring, energy physics, nuclear energy, computational genomics, climate modeling. It also should have commercial value in inter-networking applications such as private enterprise networks, multimedia distributed databases, and entertainment industry network applications. Due to its network independence, it also should be applicable as a transport over link layer technologies such as Ethernet-based passive optical networks.