Scientific communities such as Astronomy, Earth Science, and High Energy Physics rely on globally-distributed high-performance computing environments to solve old problems and make new discoveries. Even though network speeds have kept pace with Moore's Law and 10 Gbps optical-based networks span global distances, Abilene measurement data shows that less than 1% of those transfers is near to or exceeds 100 Mbps! Scientists, and their support staff, should expect much higher transfers rates (e.g., less than one minute to transfer a DVD between any two points in the global Internet). Once expectations begin to rise, tools and procedures must be readily available to allow stakeholders to find, report, and fix problems anywhere along the end-to-end path. This combination of tools, documentation, and experience will help scientists continue to achieve their research goals.
Problem Description [HTML] | Full Paper [PDF]
As Ethernet speeds have increased, there has been a widening gap between ability
of the novice to fully use bandwidth capabilities, compared to that of network
capabilites achieved by users who have had their network tuned by an expert ('Wizard"). It takes a network Wizard
to make the best possible use of an advanced network. [Diagram c/o Matt Mathis Papers & Citations]
Bridging the Gap: End-to-End Networking for Landmark Applications
Report [HTML] -- [PDF]
August 30-31, 2005
Ann Arbor, Michigan
Purpose and Goals:
Networking problems are growing and must be addressed. Some of these problems are due to scientific communities and applications using far less network power than they could, given the tuning done by network "wizards." This workshop aims to bridge this "wizard gap." It differs from other similar efforts by uniting four groups involved in network applications who depend on each other but rarely come together. In this workshop these groups will collectively discuss and coordinate approaches for improved network performance for scientific endeavors.
A goal of this workshop was to generate more productive interactions between 1) network experts or "wizards," 2) members of scientific research communities who use demanding, network-driven applications, 3) application developers and maintainers, and 4) campus network engineers who support local campus network infrastructures. Members of these various groups will explore how their distinct perspectives and concerns prompt, at times, different and, at times, overlapping orientations to problems.
The workshop resulted in an action plan for moving forward on the following
two issues related to cooperation and coordination:
Participants for this workshop were evenly distributed across the four groups. Participants discussed their respective priorities and applied them to problem-based scenarios. They explored "solutions" to scenario problems and identified the human and technical strategies required for realizing "good solutions" and for overcoming obstacles. Wizards provided commentaries. Time was devoted, as well, to Open Clinic question & answer sessions.
- Strategies for having campus and regional network operators assume and apply some of wizards' expertise and for improving their links to wizards and researchers.
- Identifying top priority problems and the processes and exchanges between groups that need to come into play in order to solve these problems.