The University of Washington is home to leading research programs conducted by internationally recognized faculty from diverse domains including oceanography, astronomy, physics, genomics, computer science, bioengineering, and climate modeling, among others. These efforts depend on high throughput data transfers between UW and remote sites, as well as between UW laboratory systems and centrally managed scalable research compute and storage systems residing in our campus network Science DMZ.

This project enhances the connectivity options of the University of Washington campus network facility dedicated exclusively to research, using a hybrid Layer-2/Layer-3 (L2/L3) approach. These enhancements replace previous ad hoc approaches to meeting the requirements of leading researchers with sustainable and scalable approaches. The enhancements are catalyzing advances in a broad range of high-impact science projects by enabling the emerging world of data-driven discovery. The UW eScience Institute provides the "intellectual infrastructure" for data-driven discovery. The enhancements generated from this project provide a critical component of the "physical infrastructure". The two together are essential for UW to remain at the leading edge as the nature of discovery evolves.

This effort directly and immediately provides a dramatic enhancement to the capabilities of a broad range of science programs involving faculty, postdoctoral fellows, research scientists, and undergraduate and graduate students. It also dramatically increases the accessibility of key UW facilities and data to scientists regionally, nationally, and internationally, and the accessibility of remote facilities and data to UW scientists.

Project Report

"Data-intensive discovery" and "Big Data" are realities defining a broad sector of today’s basic and applied research environments. Researchers (inclusive of faculty, postdoctoral fellows, research scientists, graduate and undergraduate students) routinely contend with larger and denser data sets created and collected at rates that far surpass the rate at which the data can be processed and analyzed. As a result, delays are commonplace between the time of data collection and the time when answers embedded in research results are brought to light. The University of Washington ("UW") used the funds from its NSF OCI1244890 award - "Enhancements to Support Data-Driven Discovery at the University of Washington" - to address network enhancements that would improve the ability of its researchers to move their data quickly and efficiently and thereby decrease the time between experimental data collection and the confirmation of research results. The NSF award funded essential network components of a 100G High-Speed Research Network ("HSRN") with Science DMZ (a high-bandwidth network path bypassing rate-limiting, centralized network security devices), a high-speed interconnect to an expanded UW-funded 40G campus network backbone, and access to centralized, scalable, collaboration storage ("lolo"). A 100G interconnect of the HSRN to the regional Internet2 node at the Pacific Northwest Gigapop in Seattle provides the needed high-bandwidth pathways to research and education network facilities (and their connected partners) regionally, nationally, and globally. Initial research use of the enhanced network in summer of 2014 showed promising results. A UW Physics research group led by Prof. R.G. Hamish Robertson plays a major role in two experiments, Project 8 and Majorana. The objective of the experiments is to study the mass of the neutrino, and whether it is its own antiparticle, in order to understand the large-scale structure and the matter-antimatter asymmetry of the universe. One of the instruments used for this work is housed in a lab in the Physics and Astronomy building on the UW campus. Data from the device is produced in bursts of 200-1500MB/s for periods of a second to a few hours and is sent directly to a server housed in the same facility. Ideally, the researchers would like to transfer the data near real-time to a compute facility at the Pacific Northwest National Labs (PNNL) located in Richland,WA, more than 200 miles away. Data transfer rates from the lab to PNNL originally were on the 50-80 MB/s (= 400-640Mbps) range meaning transfers could take up to several hours (~4-6) to complete. Using UW-IT’s lolo Collaboration storage service and the High Speed Research Network (HSRN) and Science DMZ, the Robertson team was able to double throughput to 170MB/s (1360Mbps = 1.36Gbps). The Robertson team is preparing a 10Gbps Ethernet service into their lab to exploit the features of the UW-IT 2014 Cyberinfrastructure Plan – 40G campus backbone, 100G HSRN, Science DMZ, and the soon-to-be-deployed research network configuration overlay which will support standardized campus access to all of these network facilities. These upgrades should allow for even faster data transfers between the experiment’s instruments and PNNL, thereby supporting the team’s research program. UW looks forward to broader, and more regular research use of these grant-funded enhanced network facilities. Departments and groups in addition to Physics who are already scheduled to take part include Oceanography, Atmospheric Sciences, Genome Sciences, Astronomy, Bioengineering, Computer Science & Engineering, as well as centralized UW scalable collaboration storage and compute facilities. This enhanced network infrastructure complements the work of the UW’s eScience Institute which provides the "intellectual infrastructure" for data-intensive discovery. Together, these resources will enable UW researchers to remain at the leading edge of their respective disciplines.

Agency
National Science Foundation (NSF)
Institute
Division of Advanced CyberInfrastructure (ACI)
Type
Standard Grant (Standard)
Application #
1244890
Program Officer
Kevin Thompson
Project Start
Project End
Budget Start
2012-10-01
Budget End
2014-09-30
Support Year
Fiscal Year
2012
Total Cost
$497,461
Indirect Cost
Name
University of Washington
Department
Type
DUNS #
City
Seattle
State
WA
Country
United States
Zip Code
98195