Project Title: “CC* Networking Infrastructure: Building HPRNet (High‐Performance Research Network) for Advancement of Data Intensive Research and Collaboration”

Sponsor: National Science Foundation

Award amount: $499,745

PI: Farzad Mashayek (Mechanical and Industrial Engineering)

Co-PI’s: Simon T. Alford (Anatomy & Cell Biology), Maxine Brown (Electronic Visualization Laboratory), Richard Cavanaugh (Physics), Cynthia Herrera Lindstrom (Academic Computing & Communications)

Main contributor to the proposal: Himanshu Sharma (Advanced Cyberinfrastructure for Education and Research)

ABSTRACT
The sizes of scientific datasets are growing exponentially across all scientific disciplines due to several factors such as improved scientific instrumentation, social media and decreasing costs of storage. To extract real value from these geographically distant datasets, researchers need to have access to these datasets at high speeds which is typically not possible with traditional campus networks. In 2009, UIC laid the foundation of a campus research network (CRN) through NSF INCUS award that provided high-performance network access to limited number of researchers. The University of Illinois at Chicago (UIC) will now be building “HPRNet”, a high-performance research network providing last mile connectivity for over 31 additional research projects. HPRNet not only improves the ongoing research productivity, but also sets the stage for future innovations and collaborations.

For HPRNet’s deployment, 13 locations are identified at UIC where 10 to 40 Gigabit uplinks to regional, national and international R&E networks will be established. HPRNet will build upon the Science DMZ model, championed by DOE’s Energy Sciences Network (ESnet), that works in concert with the current campus research network (CRN) and a special data storage system known as Data Transfer Node (DTN) to deliver high-performance and reliable network paths for data-intensive applications, including high-volume bulk data transfer, remote experiment and/or instrumentation control, cloud computing, data-mining and advanced visualization. This infrastructure will act as a “Data Expressway” at UIC, providing connectivity to Research & Education networks (R&E) to foster scientific collaboration and discovery.


Project Title: “MRI: Acquisition of SABER: Shared Analytics and Big-data Enterprise Resource”

Sponsor: National Science Foundation

Award amount: $599,536

PI: Farzad Mashayek (Mechanical and Industrial Engineering)

Co-PI’s: Philip S. Yu (Computer Science), Fatemeh Khalili-Araghi (Physics), Ao Ma (Bioengineering), Stefan Green (DNA Services Facility), Maxine Brown (Electronic Visualization Laboratory), Suresh K. Aggarwal (Mechanical and Industrial Engineering), Max Berkelhammer (Earth and Environmental Sciences), Michael E. Johnson (Center for Biomolecular Sciences), Bing Liu (Computer Science), Sohail Murad (Illinois Institute of Technology), David P. Nicholls (Department of Mathematics, Statistics, and Computer Science), Hyowon Park (Physics), Thomas J. Royston (Bioengineering), Nebiyou Yonas Tilahun (Urban Planning and Policy)

ABSTRACT
Computation has grown ubiquitous across all scientific disciplines and research education in the past few decades. Compared to its vast research portfolio, UIC has always had disproportionately inadequate central computing resources. In 2013, to bridge this widening gap, UIC made a significant investment in UIC’s first “condo” HPC (High Performance Computing) cluster called EXTREME. Research Computing group, under the leadership of Mr. Himanshu Sharma, has been managing this cluster for about 250 users across 5 colleges and 20 research groups since its inception. Research Computing at UIC is a highly collaborative initiative with representation and initial investments from 5 colleges, the Vice Chancellor for Research and the Provost.

Despite adding 25% new capacity in its first year, EXTREME is fully committed to individual groups with no scope for further expansion. Due to the high cost of entry, absence of a fees-for-service model and architectural differences barring Big Data researchers, a group of UIC users, with extensive help from Mr. Sharma, decided to pursue federal funding to build SABER: A Shared Analytics and Big Data Enterprise Resource. SABER’s architecture will allow all of UIC researchers to utilize the same cluster with minimal application changes for both HPC and MapReduce (Big Data) type workloads. This groundbreaking architecture, originally proposed by Mr. Sharma, will provide a reference architecture for other institutions who are currently creating silo clusters for HPC and Big Data. SABER will be an open-access and fees-for-service cluster providing computational resources to researchers from 9 colleges in a broad range of computational disciplines. SABER will create more opportunities for inter- and multi-disciplinary research to advance important national initiatives such as National Strategic Computing, Precision Medicine and BRAIN.