Enabling Research with Infrastructure: From Alzheimer’s to Tinnitus to Big Bird Ardoth Hassler Associate Vice President University Information Services With Collaborators: Steve Moore, Director Research Technologies Scott Allen, Director Network Services Clay Shields, Professor, Computer Science Slide 1 Slide 1 Overview • CC-NIE – Network Infrastructure: Enabling Big-Data Science Collaboration at Georgetown • CC*IIE – NWIRED - Network Innovation for Research and Education at Georgetown: Science DMZ and Cloud Services • CC*DNI – Engineer: CRaNE – Cyberinfrastructure Research Network Engineer (pending) Scientific Research – Over 22 use cases from researchers from both Main Campus and the Medical Center were cited in these proposals. Slide 2 CC-NIE Award: Network Infrastructure: Enabling Big-Data Science Collaboration • • • Georgetown upgraded high-end networking across Hilltop campuses Improvements enabled end-to-end 10 Gbps connectivity among GU's science buildings, campus backbone, off-campus data center, Mid-Atlantic Crossroads (MAX) Gigapop, and Internet2. 1Gbps or 10Gbps connections were provided to select departmental locations including: Programs • Biology • Chemistry • Computer Science • Medicine • Neuroscience • Pharmacology • Physics Centers • Center for Brain Plasticity • Center for Functional and Molecular Imaging • Georgetown-Howard Center for Clinical and Translational Science • Innovation Center for Biomedical Informatics • Lombardi Cancer Center Shared Resources • Universal Protein Resource • Others Slide 3 Enhanced Research Infrastructure Slide 4 CC*IIE: Create Science a DMZ NWIRED - Network Innovation for Research and Education at Georgetown: Science DMZ and Cloud Services • Goals/Plans: – Provide a friction-free, high-speed network for Big Science and Big Data – Create a network for scientific collaboration internally and with external researchers – Leverage GU's extant plan for Multi-Processing Label Switching deployment on campus Slide 5 CC*IIE: Create Science a DMZ Science DMZ: • Creates “Friction Free” IP networks on and off campus • Leverages GU’s10Gbps network enhancements on campus and to Internet2 (via CAAREN) • Securely extends the network into cloud services via encrypted Layer-2/3 connection at GU’s Virginia Data Center (VDC) • Benefits Access to: – HPC and Storage at GU’s at Laurel Data Center (LDC) – Services at VDC including access to Amazon Web Services, Internet2 Net+, and compute and storage resources at National Labs, XSEDE, et al Slide 6 GU’s Science DMZ Slide 7 R Alzheimer’s Research e s e a r c h GU’s 10 Gbps network upgrade allowed Dr. Federoff’s team the ability to store, manage, and analyze over 16 TBs of data from blood tests from 525 participants, age 70, over a 5-year observational period using and validating 10 plasma phospholipid biomarkers. The results identified a potential blood test for cognitive decline that is now moving to commercialization. Slide 8 R e s e a r c h Georgetown and Howard universities received notification of a $27 million Clinical and Translational Science Award (CTSA) from the NIH. The Georgetown-Howard Universities Center for Clinical and Translational Science (GHUCCTS) is a multi-institutional consortium of medical research institutions forged from a desire to promote clinical research and translational science. The members of the GHUCCTS consortium are: Georgetown University (GU), Howard University (HU), MedStar Health Research Institute (MHRI), Oak Ridge National Laboratory (ORNL), and the Washington Veteran’s Affairs Medical Center (VAMC). Through multiple partnerships and collaborations among the member institutions, GHUCCTS is transforming clinical research and translational science in order to bring new scientific advances to health care. - See more at: http://georgetownhowardctsa.org/aboutus#sthash.IY5n8cRL.dpuf Slide 9 Next Access to High-Scale Research Data - CTSA Governance Federation Workflows Compliance Now • Patient engagement Research Network Available Patients’ Records: 3.4 Million • Common IRB • BAA & DUA • Dashboard Pre-CTSA No Access GU MedStar HUH VA DC •EHR •LAB •Registries •Research •EHR •LAB •Registries •Research •EHR •Lab •No Std. •Project based •EHR • Cohort Discovery • BI Services • i2b2 GU MedStar HUH VA DC •EHR •LAB •Registries •Research •EHR •LAB •Registries •Research •Project based •EHR •i2b2 •Project based •EHR • • • SHRINE CTSAs Friction Free research • • Network • • • • • Common IRB BAA & DUA Dashboard CTMS Coalitions Team science Underserved Pop • Cohort Discovery • BI Services • i2b2 GU MedStar HUH •EHR •LAB •Registries •Research •EHR •LAB •Registries •Research •EHR •LAB •Registries •Research VA DC •Project based •EHR Hybrid Cloud Computing solution for research Built on UIS LDC & AWS’s Virtual Private Cloud Architecture • Science as Service • Elastic environment • Storage and computing at scale Isolated AWS resources Subnets NAT Router VPN Gateway Internet Secure VPN Connection over the Internet GU 10 Gig VPC Amazon Classic Web Services Cloud Investigator s Desktop 10 Gig LAN Source: Amazon Web Services VPC Architecture Globus Genomics Project: Leveraging High Speed Bandwidth at GU: • Globus Genomics (GG): ICBI and Genomics Shared Resource at Georgetown University Medical Center initiated a joined pilot project with the University of Chicago. The project is introducing and implementing high efficiency tools and capabilities to transfer large data sets to the cloud, and use pipelines and software applications developed by the Center. – Data Management SaaS – Next-gen sequence analysis pipelines – Flexible, scalable, cost effective Globus Genomics Pilot Current Projects 8 Users 6 R e s e a r c h Neuroscientists Discover Brain Abnormalities for Tinnitus and Chronic Pain GUMC and Germany’s Technische Universität München have uncovered the brain malady responsible for tinnitus and for chronic pain. The scientists describe how the neural mechanisms that normally “gate” or control noise and pain signals can become dysfunctional, leading to a chronic perception of these sensations. They traced the flow of these signals through the brain and showed where “circuit breakers” should be working — but aren’t. https://gumc.georgetown.edu/news/NeuroscientistsUncover-Brain-Abnormalities-Responsible-forTinnitus-and-Chronic-Pain Josef Rauschecker, PhD, DSc Slide 14 Measureable Outcomes Center for Brain Plasticity and Recovery 454 MB Brain Image File Before CC-NIE: 39 seconds to upload to Data Center After CC-NIE: 19 seconds http://cbpr.georgetown.edu/ Slide 15 Measureable Outcomes Children’s Digital Media Center: Understanding Digital Media’s Influence on Children Sandra L. Calvert, Ph.D., is Professor of Psychology at Georgetown University (G.U.) and a co-founder and the Director of the Children’s Digital Media Center a multiinstitution interdisciplinary research center that has received funding primarily from four collaborative grants from the National Science Foundation. Dr. Calvert’s lab is connected to the 10 GBs campus network. She wrote: “Just wanted to thank you for getting my office and Center into the digital age!” - Sandy http://cdmc.georgetown.edu/ Slide 16 Approaches: Two Examples • AWS – Connected to Amazon Cloud using site-to-site VPN – Proteomics Information Resource Center was satisfied with test results – More to come with Science DMZ • Results of Prior Work for Notice for Prospective PIs Networking Infrastructure Enhancements: Georgetown University is the recipient of NSF CC-NIE award #1340971, Enabling Big-Data Science Collaboration at Georgetown (2013, $379,018) and CC*IIE award #1440743, NWIRED - Network Innovation for Research and Education at Georgetown (2014, $425,373). These awards enabled campus network infrastructure upgrades to 10 Gbps with 1 Gbps uplinks for most researchers, increased the speed of our connection to Internet2, and enabled implementation of a Science DMZ (in progress). When completed in 2015, the Science DMZ portion of the network will be optimized for high performance scientific applications rather than for general purpose business systems. [Ardoth A. Hassler, PI] Slide 17 Summary • What’s Ahead – BIG Data • AvesTerra infrastructure: force migration, HIV/AIDS – http://avesterra.georgetown.edu/ • McCourt School of Public Policy Massive Data Institute – http://mccourt.georgetown.edu/massive-data-institute/research – Much more… • Computer Science Collaborations • Neuroscience is GU’s most externally funded research • Tinnitus Research is one of many projects that will benefit from a CC*DNI award Results of CC-NIE and CC*IIE awards are enabling research at Georgetown Slide 18 Questions? Contact • Ardoth Hassler hasslera@georgetown.edu • Steve Moore moores@georgetown.edu • Scott Allen Scott.Allen@georgetown.edu • Clay Shields clay@cs.georgetown.edu Slide 19