KAREN What you can do with an advanced

KAREN What you can do with an advanced research and education network!

Introductions o John and Sam n n n o We do not know your science We want to facilitate discussion This is an opportunity to report back to REANNZ on issues and barriers Who are you? 2

Today’s Plan o o o Introduction Collaboration – now and in the future Lunch Tools Capability Development Wrap up 3

Introduction o Motivation n o o Research Networks E-Research n o A paradigm shift What is it? International trends n Examples 4

The New Research Paradigm Credit: GEANT 2 5

Case Study: Serious Disease Genes Revealed o o o Wellcome Trust Case Control Consortium 50 research groups 200 scientists DNA from 17, 000 patients 15, 000 polymorphic markers Learned more in 12 months than last 15 years 6

Case Study: Functional MRI (f. MRI) Data Center o o o Online repository of neuroimaging data A typical study comprises n 3 groups n 20 subjects/group n 5 runs/subject n 300 volumes/run n 90, 000 volumes, 60 GB raw data n 1. 2 million files processed 100 s of such studies in total Credit Ian Foster, University of Chicago 7

www. fmridc. org 8

Global R&E Network Pathways DISCLAIMER - This network map was a best estimate of expected connectivity for 2005, several changes in connectivity and planned connectivity have happened since it was created Credit: John Silvester, USC, Chair CENIC 9

Kiwi Advanced Research and Education Network Credit: KAREN. http: //www. karen. net. nz 10

KAREN o o o o Went live Dec 2006 10 Gb/s NZ Backbone $40 million, Government Funding $5 million Capability Build Programme Linking all 8 Universities and all 9 Crown Research Institutes, + National Library ~622 Mb/s link to US ~133 Mb/s link to Australia Credit: KAREN. http: //www. karen. net. nz 11

Advanced Research and Education Networks (ARENs) Credit: GEANT 2 12

What is e-Research? o o o Collaboration Access to and management of data and knowledge Advanced computing methods Shared resources New research techniques 13

Characterising e-Research Characteristic Traditional Research E-Research Participants Individual researcher or small local research team Diversely skilled, distributed research team Data Locally generated, stored and accessible Generated, stored and accessible from distributed locations Computation and Instrumentation Batch compute jobs or jobs run Large-scale, or on demand on researcher’s own computers computation or access to or research instruments shared instruments Networking Not reliant on networks Reliant on research networks and middleware Dissemination of Research Via print publications or conference presentations Via web sites and specialized web portals Credit: Bill Appelbe and David Bannon, Victorian Partnership for Advanced Computing. e. Research: Paradigm Shift or Propaganda? http: //www. jrpit. acs. org. au/jrpit/JRPITVolumes/JRPIT 39. 2. 83. pdf 14

Discussion o o Where does your research fit into this characterisation of traditional research and e-research? How does this compare with the research that you were doing 5 years ago? 15

Current Environment - Set of Tools video conference web sites experiment instrument scientist email data storage HPC analysis Credit: Be. STGrid. http: //www. bestgrid. org 16

Future Environment Research Collaboratories scientist Grid Middleware scientist instrument web portals experiment messaging scientist Credit: Be. STGrid. http: //www. bestgrid. org HPC data storage video conference analysis 17

The Researcher’s View o Why do I care? n n n o What’s in it for me? n n n o New collaborative opportunities New funding opportunities NZ competitiveness Key resource is often somewhere else More data, more tools Collaborating with the best How do I get involved? n Move from silo to GRID Credit: Be. STGrid. http: //www. bestgrid. org 18

Example e-Research Projects o o o Bio. Co. RE SCOOP SEEK/Eco. Grid 19

20

Bio. Co. RE o o o o Seamlessly access local and remote technology Co-author papers Access high performance computing Share molecular visualisations Chat room Lab book Notifications, etc. http: //www. ks. uiuc. edu/Research/biocore/ 21

The Control Panel 22

Projects 23

Project Summary o Review n n n State of recent job submissions Who is logged in What tasks members are working on Recent discussion topics Recent files added to Bio. FS 24

Project Status o See n n o Modify n o Current work Future work Schedule of upcoming tasks Display n Current task 25

Publishing VMD Sessions 26

Configuring NAMD Simulations 27

Job Management o A Grid Portal n n o Submit web form Monitor progress Bio. Co. RE n n Obtains resources Moves files Executes jobs Places results 28

Message Board 29

Lab Book 30

Website Library 31

Bio. Co. RE File System 32

SURA Coastal Ocean Observing and Predicting Programme 33

SCOOP o o o Promote effective and rapid fusion of observed oceanic data with numerical models Facilitate the rapid dissemination of information to operational, scientific, and public or private users http: //scoop. sura. org/ 34

SCOOP Goals o Create an open access, distributed lab for oceanography by: n n n Supporting data standards development and implementation Demonstrating benefits/added value of diverse communities moving to common standards for info exchange Creating an environmental prediction system –a research tool that can also support relevant agency decision-making to improve society 35

Real-Time Ensemble Prediction Each forecast wind field is used as input for numerical predictions of storm surge of waverelevant and availableand disseminated fields. are visualized observations are Hurricane warnings Results from each analysis the NOAA National Hurricane are then For resultsandthe issued by. Because each individual element in this The verification, allof the predictions in the ensemble. Center in ensemble of surgebe readilypredictions involves adecision-support (NHC) are used to compared incorporated into wind fields. aggregated and create an ensemble of forecast which show the a form that can analysis. Results include maps that provides a for and wave with predictions, numerical calculation measure of accuracystreet level set of forecast winds Each of theseof emergency with andon a large supercomputer probability wind fields represents a quality for real-time by could take many hoursplausible detail. tools usedthat inundation response personnel. the predictions. cluster, they are farmed out to the available computational resources over the entire region of interest for several days into the future. within the distributed network. 36

Distributed Facility for Coastal Prediction wind forecasts water level model wave watch Open. IOOS model data 37

Science Environment for Ecological Knowledge o Aims to extend ecological and biodiversity research capabilities by fundamentally improving how researchers: n n n o gain global access to ecological data and information find and use distributed computational services exercise powerful new methods for capturing, reproducing & analysing data http: //seek. ecoinformatics. org/ 38

SEEK’s Integrated Systems o o o Eco. Grid n Next generation internet architecture enables data storage, sharing, access and analysis Semantic Mediation System n Advanced reasoning system determines if data and analytical components can be automatically used in a selected workflow Analysis and Modeling System n Ecologists design, modify and incorporate analyses to compose new workflows and models in a visual, automated environment 39

Eco. Grid o o o o Seamless access to and manipulation of data and metadata stored at different nodes Authentication via single sign-on Web services for executing analytical pipelines Registry of data and compute nodes Rapid ingest of new data sources as well as decades of legacy data Extensible relevant metadata based on the Ecological Metadata Language Data replication provides fault tolerance, disaster recovery and load balancing 40

Kepler Workflow Tool o Example of the 'R' system in a Kepler workflow 41

Things to take away o o o The research lifecycle is changing – an evolution rather than a sea-change Bigger and more complex problems require new methodologies and relationships Policy and funding are increasingly dictating collaboration Advanced networks are essential It’s more about data than technology Many social and organisational factors 42

A Final Message Credit: GEANT 2 43