University of Missouri Internet2 Spring Member Meeting April 24, 2012 Developing a 100G TestBed for Life Science Collaborations Taking advantage of existing UM/SURA dark fiber to create a research 100G pathway from St Louis to Kansas City via Columbia along Interstate 70 Using InCommon Federated Identities for authentication via Shibboleth; authorizations occur via an autonomous Entitlement Server to provide fine-grained authorizations to/from service providers Developing distributed resource sharing by mapping needs to resources eligible for assignment depending on geography and resource availability At 100G some distributed computing latency issues can likely be overcome Problem Set High-Throughput Sequencing is producing enormous quantities of data; needing a storage infrastructure as a private cloud Need to provision collecting, analyzing and using resources according to demand and includes the processing applications being net-aware Develop security, measurement and analysis tools to efficiently run at 100G across a regional multi-cluster environment using OpenFlow and other specialized protocols Where Does the Data Come From? A High-Volume EST Pipeline For Discovery High throughput DNA sequencing at MU DNA Core Facility. Swine female reproductive tissues and embryos are removed at various times of gestation. Gene annotations quickly obtained by access to other databases through the MU Internet2 high speed network. Improved efficiency, quality and profitability is the goal. Sequence data analyzed at MU on high-speed systems. Iterate microarray & other experiments to focus on gene discovery. Patterns of gene expression analyzed with microarrays to reveal mechanisms that contribute to reproduction efficiency. KC I70 SL Col Big Data/Big Science Collaboratory The GPN Network University of Missouri System Grant Writers Professional Development Sessions UM InterCampusNetwork with 100G Pathway along I70 HtSeq LSC Internet2 100G UM Portion of MOREnet 100G Multi-Site Sharing HPC Infiniband Network HPC Infiniband Network Mgmt HPC Storage SMP Servers Linux Cluster GPGPUs Mgmt HPC Storage Machine Room GbE Network Managem ent Central Administra tion Monitoring File Mgmt Availabilit y Data Migration Replicatio n Backup Research Data Store H C R U6 H C R U6 IBM CLOUD IBM H C R U6 IBM H C R U6 IBM OpenFlow And Other CLOUD Protocols Instruments (Core Service) Visualization & Display Instruments (Medical) Site A GPGPUs Protocols CIFS NFS HTTP FTP SCP H C R U6 Managem ent Central Administra tion Monitoring File Mgmt H C R U6 Availabilit y Data Migration Replicatio n Backup Research Data Store Campus GbE Network Lab (Research & Clinical) Linux Cluster Machine Room GbE Network Protocols CIFS NFS HTTP FTP SCP LOGIN Next-Gen IBM LOGIN IBM H C R U6 IBM H C R U6 IBM Visualization & Display Campus GbE Network Instrument (Research) Lab (Research & Clinical) Instruments (Core Service) Instruments (Medical) Site B Instrument (Research) OpenFlow at the packet level Controller OpenFlow-enabled Commercial Switch Normal Software Normal Datapath PC Secure Channel Flow Table Analysis Engines User Storage Cloud NetProcessors NetStorage Adapted from: The Stanford Clean Slate Program http://cleanslate.stanford.edu Collaborative Framework 11 Bridging the Gaps (Some are very Large) • Authenticate • Provider (InCommon) Analysis Sharing Tools Nets • Processors Apps • Data Administrative User WAYF 2 4 5 Identity Provider 14 5 Command Service Provider 1 Credentials Handle Service Identity Directory 3 User Handle 6 SHIRE 6 Attribute Authority Handle 7 SHAR Credentials 8 Attributes 12 Entitlement Client App Entitlement Server ES DB 13 YES/NO 9 VO Entitlement Command Using Middleware Tools for VO Collaboration Resource Reso urce Manager Handle Command 10 Entitlement Server YES/NO 11 User Simplified Design 1: request by URL or command 2 Entitlement Server 5 4 uses public key encryption for authentication and privacy Identity Provider 3 Service Provider 6 Page or computational results Getting Authenticated If you belong to a GPN member organization, but do not see your institution in the list, please contact your local GPN representative to request help in authenticating in this environment. Entering the VO Environment 2/24/2008 16 And the story continues … More Data, Resources, People & Knowledge UMBC http://umbc.rnet.missouri.edu 02/14/2007 Grant Writers Professional Development Sessions 18