CC-NIE Network Infrastructure: Enabling data-driven research

Principal Investigator: Jay Aikat

Funding Agency: National Science Foundation

Agency Number: OCI-1245783

Abstract:

This project identifies science labs at the University of North Carolina Chapel Hill with large data-intensive research projects on campus that would benefit most from improvements in network infrastructure. The project focuses on network enhancements by replacing networking equipment, both building links and access layer, related to these research labs on campus.

The enhancements consist of upgrading the existing 100 Mb-only switches in the affected departments with gigabit Ethernet switches providing multi-gigabit links back to upgraded building aggregation switches that, in turn, connect to the existing network backbone at 10Gbps. This provides 10Gbps connectivity from each building to the campus research computing services, as well as to Internet2 resources. The enhancements also include a monitoring machine equipped with a 10Gbps DAG monitoring card to collect traffic header traces at the ingress-egress routers at the edge of the campus network. These enhancements will enable better and faster access to the large amounts of data transferred regularly by these labs, thus enhancing data management in their respective research projects.

The research labs directly and significantly impacted by this project are involved in the following research projects:

* Biological Sciences: producing mathematical models of animal flight (e.g., hummingbird, butterfly), using recordings of high-speed and high-resolution video from locations throughout the campus.

* Biological Sciences: studying the evolution and genetics of de novo genes; obtaining the data, particularly the raw data, demands several minutes to several hours of uninterrupted downloads.

* Geological Sciences: collecting, aggregating, and moving data from volcanic and seismic monitors (some in real-time), real time analysis of ongoing geologic processing.

* Environmental Science and Engineering: Lattice Boltzmann modeling that requires large lattice sizes (8003) that translate to state storage of 8.7-95.3GB per state point.

* Computer Science: transferring large network traffic datasets captured at the main campus link to clusters of data processing machines in the lab for use in network experiments.

This project additionally has direct impact on the undergraduate and graduate curricula in the various departments represented by these research labs, broadening the participation of student researchers. These research labs continue their investment in promoting undergraduate research by involving undergraduate students, including actively promoting women and minorities in their groups.