The Cyberinfrastructure Movement and the Potential for
Revolutionizing Science/Engineering Research and
EducationDaniel E. Atkins
[email protected] of Information & Department of EECS
University of Michigan, U.S.A.
Oxford Internet Institute, July 5, 2004
D. E. Atkins • University of Michigan • [email protected] 2
Home Land Securityhttp://web.calit2.net/RiskReduction/index.html
Converging Streams of Activity
Science-driven pilots (not using above labels)
E-science
Collaboratories
Cyberscience
GRIDS (broadly defined)
2nd Editionwww.mkp.com/grid2
ACLS-MellonCI for Humanities
IT & Future of Higher Education
“a new age has dawned in scientific and engineering research, pushed by continuing progress in computing, information, and communication technology, and pulled by the expanding complexity, scope, and scale of today’s challenges. The capacity of this technology has crossed thresholds that now make possible a comprehensive “cyberinfrastructure” on which to build new types of scientific and engineering knowledge environments and organizations and to pursue research in new ways and with increased efficacy.”
http://www.cise.nsf.gov/sci/reports/toc.cfm
NSF Blue Ribbon Panel on Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 4
Cyberinfrastructure-enabled Knowledge Communities* (CKCs)
Computation, Storage, Communication and Interface
Technologies
Cyber-infrastructure:Equipment, Software, People, Institutions
Virtual teams, communities, organizations, knowledge
communities/environments/ecologies
* or perhaps “Organizations”
D. E. Atkins • University of Michigan • [email protected] 5
“Knowledge Communities”This phase is shorthand for a group of people working together to create, disseminate, use and/or preserve knowledge.
I do not mean a static collection of knowledge (as in “a body of knowledge”) -- I mean people engaged in knowledge-based activities. (”Knowing Communities”)
Not one huge knowledge community but rather many specialized communities, often with overlapping membership. Individual role may vary among different knowledge communities.
research, scholarship
teaching, learning
D. E. Atkins • University of Michigan • [email protected] 6
Some Names for CKCs
Co-laboratory, Collaboratory
Grid Community
e-X Community (as in e-science)
Cyber-X Community (as in cyberscience)
Community Gateways or Portals
Virtual Community, Virtual Organizations
D. E. Atkins • University of Michigan • [email protected] 7
Big IdeasGlobal cyberinfrastructure can become a platform for routine, effective, computationally supported, distance-independent activities of knowledge communities. (Goal is not to eliminate same time and place collaboration, but rather to augment it, and use it in more effective ways).Research communities are creating functionally complete virtual communities that are absolutely necessary for their next decade of research aspirations.Cyberinfrastructure offers new options for what is done, how it is done, and who participates.We now have the opportunity (and responsibility) to help make it real.
D. E. Atkins • University of Michigan • [email protected] 8
Computation
Content
Interaction
Cyberinfrastructure to
enable knowledge
environments to
revolutionize science
and engineering
education and
research
Vision & requirements
from the frontiers of
science and engineering
research
The Push The PullThe Goal of an ACP
*ACP = Advanced Cyberinfrastructure Program
*
But the implications go way beyond science and engineering research.
Push and Pull for CKCs
D. E. Atkins • University of Michigan • [email protected] 9
Dimensions of CI R&D
usefulness
D. E. Atkins • University of Michigan • [email protected] 11
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 12
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 13
D. E. Atkins • University of Michigan • [email protected] 14
Translight Consortia
D. E. Atkins • University of Michigan • [email protected] 15
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 16
NSF Middleware Initiative
http://www.nsf-middleware.org/
D. E. Atkins • University of Michigan • [email protected] 17
Middleware: Globus and Grid-Web Services
Convergence
Web Services Messaging, Security, Etc.
Open Grid Services Infrastructure
Domain-Specific Services
Core Services
ProgramExecution Data Services
From Ian Foster
D. E. Atkins • University of Michigan • [email protected] 18
Core MiddlewareIdentity and Identifiers – namespaces, identifier crosswalks, real world levels of assurance, etc.Authentication – campus technologies and policies, interrealm interoperability via PKI, Kerberos, etc.Directories – enterprise directory services architectures and tools, standard objectclasses, interrealm and registry servicesAuthorization – permissions and access controls, delegation, privacy management, etc.Integration Activities – open management tools, application of virtual, federated and hierarchical trust, enabling common applications with core middleware
D. E. Atkins • University of Michigan • [email protected] 19
Mellon Foundation Higher-Ed Open Source Projects
Project DescriptionPubcookie Authentication system including “single sign on.”OKI Open standards for sharing digital objects.PKI Inter-institutional public key infrastructure.ePortfolio Electronic portfolio tools for higher ed.uPortal Web portal development software.AAM Tool for managing course assignments & tests.LionShare Authenticated P2P networks for legit file sharing.SAKAI Feature-rich course management system.OCW Free worldwide access to educational content.VUE Visual understanding environment for digital content.Chandler Personal information manager for higher ed.DSpace Digital repository system federation.Fedora Flexible Extensible Digital Object Repository Architecture.
http://rit.mellon.org/twiki/bin/view/Main/PubcookieTwiki
D. E. Atkins • University of Michigan • [email protected] 20
Open Middleware Infrastructure Institute
D. E. Atkins • University of Michigan • [email protected] 21
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 22
Computation-Simulation
• Capability not just capacity: technology, policy, tools.• Still need some center-based leadership,super computers.• On-demand supercomputing, not just batch.
Report of the High-end Computing
Revitalization Task Force (5-04)
D. E. Atkins • University of Michigan • [email protected] 23
Japanese Earth Simulation Center
D. E. Atkins • University of Michigan • [email protected] 24
Top 5 Supercomputers
From http://www.top500.org/list/2003/11/
D. E. Atkins • University of Michigan • [email protected] 25
Virginia Tech Terascale Cluster (1,100 Mac G5s)
http://computing.vt.edu/research_computing/terascale/
D. E. Atkins • University of Michigan • [email protected] 26
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 27
Some Implications for the “DL Research” Community
Onward to the GII - ubiquitous knowledge environments and information ether.
Increased variety and scale of information.
New scholarly communication systems
Reducing “participation overload.”
The openness movement.
Persistence of access - digital preservation.
See http://www.sis.pitt.edu/~dlwkshop/
D. E. Atkins • University of Michigan • [email protected] 28
Open source software and its communities
Open standards
Open content; open knowledge. Digital repositories.
Open alliances for creating middleware
Open intellectual properties more broadly
Keeping the Internet architecture open
University as counterbalance to overly restrictive access (”rights management”)
Openness Movement
D. E. Atkins • University of Michigan • [email protected] 29
http://www.si.umich.edu/digarch/http://www.digitalpreservation.gov/index.php
Research Issues!
Includes “Preserving Our Digital Heritage” report & “Time” report.
Research Challenges in Digital Archiving and Long-Term Preservation
D. E. Atkins • University of Michigan • [email protected] 30
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 31
NEESgrid Earthquake Engineering Collaboratory
www.neesgrid.org
D. E. Atkins • University of Michigan • [email protected] 32
Reconstructions of Dendritic Spines by High
Performance Parallel Electron Tomography
This 3MeV Electron Microscope in Osaka Japan is the Biggest in World
The Spines on these nerve cell dendrites are where connections are rapidly made and
unmade between cells in the brain.
The Initial Facility at Sondrestrom, Greenland
The University of Michigan Upper Atmospheric Research Collaboratory (UARC)
D. E. Atkins • University of Michigan • [email protected] 34
Embedded Sensors: R&D & Use
Ocean Research Interactive Observatory Networks
http://www.coreocean.org/Dev2Go.web?Anchor=orion_home_page&rnd=17953http://www.nsf.gov/bio/neon/start.htm
National Ecological Observatory Network (NEON)
http://www.cens.ucla.edu/index.html
D. E. Atkins • University of Michigan • [email protected] 35
D. E. Atkins • University of Michigan • [email protected] 36
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 37
Electronic Visualization Labhttp://www.evl.uic.edu
Tele-Immersive Collaboration in the CAVE Research Network
D. E. Atkins • University of Michigan • [email protected] 38
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 39
Time-Space Collaboration
Physically together...
Drop in lab, physical library,
museum
Audio, chat, video conference, group
applications
Email, threaded-discussions, shared
files...
Same
Same
Different
Different
Time
Place
distance matters beyond being there
D. E. Atkins • University of Michigan • [email protected] 40
Some examples of experimental CKCs
The Initial Facility at Sondrestrom, Greenland
The University of Michigan Upper Atmospheric Research Collaboratory (UARC)
D. E. Atkins • University of Michigan • [email protected] 42
UARC Interface
team chat
dynamic work
rooms
Real-time instruments
Archival data Journals
computational models
anno
tatio
nSe
ssio
n re
play
Evolved into a Network of Instruments (one global instrument)
D. E. Atkins • University of Michigan • [email protected] 44
UARC Patterns of Communication
1998Smithsonian Science
Award
http://crew.umich.edu/
D. E. Atkins • University of Michigan • [email protected] 45
Vignettes: UARC/SPARCShared, tele-instruments & expertise.
Rapid response, opportunistic campaigns.
Multi-eyes, complementary expertise.
Isolated instruments became a global instrument chain.
Cross-mentoring/training.
New & earlier opportunities/exposure for grad students.
D. E. Atkins • University of Michigan • [email protected] 46
Enhanced participation. Legitimate peripheral participation.Support for authentic, inquiry-based learning at UG and pre-college level.Distributed workshops for post-campaign data analysis.Session re-play for delayed participation.Data-theory closure.A “living specification” to stretch vision of possibilities.
D. E. Atkins • University of Michigan • [email protected] 47
Cyberinfrastructure is a First-Class Tool for Science
D. E. Atkins • University of Michigan • [email protected] 48
The Information Technology Needs of the NeuroscienceCommunity Provide Major Challenges for Tomorrows
Information Technology Infrastructure
• BRAIN RESEARCH of the Future will be conducted in a DISTRIBUTED ENVIRONMENT
• The NSF PACI’s and Emergence of new Information Technologies have catalyzed this change
• Projects to Federate Neuroscience Data, Build Data and Computational Grids and Telescience Tools are creating a Cyber Infrastructure that will Enable New Science in this and other disciplines
Enable new understanding of the brain by linking data about macroscopic brain function
to its molecular and cellular underpinnings
http://www.nbirn.net/
Bioinformatics Research Network (BIRN)
D. E. Atkins • University of Michigan • [email protected] 49
Crab Nebula in 4 spectral regions:X-ray, optical, infrared, radio
http://www.us-vo.org/
Virtual Observatory Prototype Produces Surprise Discovery. Early demo project identifies new brown dwarf.
http://www.us-vo.org/news/brown-dwarf.html
D. E. Atkins • University of Michigan • [email protected] 50
Broader Implications
D. E. Atkins • University of Michigan • [email protected] 51
Cyberinfrastructure for Multi-use
New ACLS Panel of CI for Humanities
D. E. Atkins • University of Michigan • [email protected] 52
Knowledge Communities Enabled by
Cyberinfrastructure
teachinglearning
serviceengagement
researchscholarship
Potential Mission Synergy: Shared-Use CKCs
CLEAR-Collaborative
Learning, Engagement, and
Research
D. E. Atkins • University of Michigan • [email protected] 53
Sloan DigitalSky Survey/Skyserver
Johns Hopkins UniversityAlex Szalay (Co-PI), Jordan Raddick
Database of 80+ million objectsSearch and Analysis ToolsProjects for Teachers
and Students
D. E. Atkins • University of Michigan • [email protected] 54
OpportunitiesUse the same data and tools
Engage students and their teachers in authentic astronomy researchBecome part of the ongoing research communityIncrease the pace of scientific discovery.
D. E. Atkins • University of Michigan • [email protected] 55
BenefitsStudents who engage in authentic research will understand the process of science
Be more likely to pursue a career in science
Grow up to be scientifically literate adults
Improve science teaching through ongoing teacher professional development and mentoring
Reconnect with science
Reduce isolation
Increase Retention
D. E. Atkins • University of Michigan • [email protected] 56
Global Examples..Global Graduate Seminar on ICT & Globalization - Cogburn, UM/SI
Global Product Design - Dutta UM/ME
D. E. Atkins • University of Michigan • [email protected] 57
CKCs -- lofty aspirations…Provide greater equity of access and participation.
Open up more experiences and increase the probability of intellectual linking across disciplinary boundaries.
Enrich the diversity of participation, perspective, ideas, experiences.
Enable sharing of resources and better amortization of unique resource/facilities.
Support existing teams and accelerate the formation of new teams, fields, disciplines.
Support rapid ad-hoc team formation to respond to unexpected emergencies.
D. E. Atkins • University of Michigan • [email protected] 58
Comments on the Research Model to Create and Apply
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 59
Creation of knowledge: basic, curiosity-driven
research
Application of knowledge
Focus on New Knowledge Creation?
Focus on Application?No
No
Yes
Yes
Edison
Bohr Pasteur
Pasteur’s Quadrant Research Model
Creating effective CKCs require PQ research models
CKCs can support PQ research models
Classic Linear Research Model
D. E. Atkins • University of Michigan • [email protected] 60
Borromean Ring Design Teams
Three symmetric, interlocking rings, no two of which are interlinked. Removing one destroys the synergy.
Users,Communities of
Practice,Organizations in
Society
Behavioral & organizational
sciences & practice
Computer & information,
science, engineering &
practiceTeams
learning together - collateral learning
D. E. Atkins • University of Michigan • [email protected] 61
CKCs R&D Approach
Human-centered
Experimental
Iterative Design
Long-term
but robust & usable
B-teams
D. E. Atkins • University of Michigan • [email protected] 62
Forces on Higher Education
Massification
Demographics
Private knowledge production
Internationalization
Changing IP regime
Competence recognition
Social fragmentation
Patterns of management
ICT
D. E. Atkins • University of Michigan • [email protected] 63
Forces on Higher Education
Massification
Demographics
Private knowledge production
Internationalization
Changing IP regime
Competence recognition
Social fragmentation
Patterns of management
ICT
Cyberinfrastructure
D. E. Atkins • University of Michigan • [email protected] 64
big and smalllocal and globalcentralized and decentralizedlearning and research/discoverybasic and applied (Pasteur’s Quadrant)giving and taking (reciprocity)producer and consumermulti roles: expert, student, observer, policy-maker
“And-And” CI-E* Organizations & Activities
*cyberinfrastructure-enable
D. E. Atkins • University of Michigan • [email protected] 65
Budget Recommendation Overview(Incremental, Recurring)
Fundamental research to advance CI
$60M
Application of CI to advance S&E research
$200M
Provision & operational CI
$660M
Information and data repositories
$200M
TOTAL $1020M
D. E. Atkins • University of Michigan • [email protected] 66
FrameworkScience and Engineering Frontiers (Frontiers): S&E opportunities to be realized using CIIntegrating Architectures (IAs): limited number of common architectures that support domain-specific applications using a common, reconfigurable set of open source software tools, technologies and services.
Computation-intensiveInformation-intensiveInstrumentation-richDesk-top
D. E. Atkins • University of Michigan • [email protected] 67
Core (Core): CI foundation including backbone networks, widely shared compute and storage facilities; education and workforce development activities; a portfolio of activities aimed at yielding new knowledge on the science of cyberinfrastructure, including its human and social dimensions and rigorous evaluation and assessment activities.
CI-enabling Research (CI-R): research investments that will create new information technology tools and resources to enrich cyberinfrastructure for the foreseeable future.
D. E. Atkins • University of Michigan • [email protected] 68
CoreEducation, Training, Community Development: Establish coordinated programmatic activities to
prepare current and future scientists and engineers to use, develop and support cyberinfrastructure;catalyze domain S&E community efforts to define unique CI-enabled research and education opportunities (coordinated with Frontier activities);take targeted action to broaden participation of underrepresented groups and organizations in CI activities.
D. E. Atkins • University of Michigan • [email protected] 69
Networks: Undertake comprehensive NSF-wide planning to:
examine current international investments in network infrastructure to support Frontier projects now and in the future;
develop network infrastructure needs assessment and gap analysis;
identify last-mile connections solutions and priorities
Informed by the above determine programmatic strategy to invest in enabling networks both Frontier and Core approaches as appropriate.
D. E. Atkins • University of Michigan • [email protected] 70
Supercomputing Platforms: Define complementary supercomputing investments to be made through Core, CI-R, and Frontier activities. To inform these investments
conduct comprehensive assessment of national supercomputing resources
identify promising interagency strategies to address supercomputing needs for open-science community
identify organizational and economic models that support centralized and/or distributed supercomputing investments
D. E. Atkins • University of Michigan • [email protected] 71
Federated Data Archives and Digital Libraries: Undertake comprehensive NSF-wide planning to:
explore the efficacy of creating a national databank of federated data archives. recognizing that data assets reside in different locations and belong to multiple domains, individuals and organizations;
assess domain-specific interests in establishing and supporting digital libraries;
identify organizational and economic models that support centralized and/or distributed data archives and digital library investments.
D. E. Atkins • University of Michigan • [email protected] 72
Science of Cyberinfrastructure
Study the impact and use of cyberinfrastructure in research and education, to develop a better understanding of its sociological, economic, technological and societal implications.
Initiate a comprehensive research, evaluation and assessment activity that seeks to develop new knowledge on the most effective strategies to design, develop and use cyberinfrastructure with emphasis on usability, accessibility and scientific utility.
Questions, Discussion