Transcript The Coming of Cyberinfrastructure

The Coming of Cyberinfrastructure
Gary M. Olson
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Cyberinfrastructure



“Infrastructure” – used since the 1920’s to refer to the
roads, bridges, rail lines, and similar public works
required for an industrial economy
“Cyberinfrastructure” – infrastructure based on
computer, information, and communication technology
used for discovery, dissemination, and preservation of
knowledge
Cyberinfrastructure – required for an information
economy
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
NSF Cyberinfrastructure Report

Blue Ribbon
commission
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Process




NSF will make cyberinfrastructure a major
initiative
Hosted a series of workshops and on-line town
meetings
New division of CISE formed
Coordinated with other federal agencies
–
NIH, DARPA, DOE, NASA
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Cyberinfrastructure: the Middle
Layer
Virtual teams, communities, organizations,
knowledge environments/ecologies
Cyber-infrastructure:
Equipment, Software, People, Institutions
Computation, Storage, Communication and
Interface Technologies
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Computation
• Capability not just capacity: technology,
policy, tools.
• Varying levels of need – mix of
commercial and special-purpose tools
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Cyberinfrastructure - more than ICT







Equipment: hardware, software, boxes, wires,…
Organizations
Policies
Processes
Reliability
Sustainability
Trained human resources to create, operate, apply.
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Adoption of cyberinfrastructure

Mix of technical and social issues
–
–
–
–
Raw technology
Standards
Utility
Network effects
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Example: Electronic Mail






1963 -- ARPA had time sharing system with messaging (J. Licklider)
1970 -- ARPANET operational w/ 4 nodes
 UCLA, UCSB, Utah, SRI
– 1971, 23 nodes
– 1973, 45 nodes
– 1976, 111 nodes
1975, more than 1000 registered e-mail users on ARPANET
1978, Usenet for non-ARPANET universities
Proliferation of networks in late 70s, early 80s
TCP/IP as networking standard --> Internet in mid 80s
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Needs for Cyberinfrastructure

Many needs articulated in CI report

But the ability to handle large amounts of data
the most widely expressed need
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Examples of needs






NEESgrid – earthquake engineering data
HIV/AIDS – large scale clinical trials
Ecology – integration in time and space
BIRN – brain images
National virtual observatory – astronomy data
High energy physics – outputs from high end
accelerators
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Science Technical
Federating
data
Issues
bases
Metadata
Access rights
Social
Organizational
for large
of data (1) ●
● amounts
●
●
●
Security
●
●
●
●
●
●
●
Access speed
●
●
Size, storage
capacity
●
●
Storage speed
●
●
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Science Technical Social Organizational
Provenance
Issues
for large
amounts
●
●
● of data● (2)
●
Standards
●
Data mining
techniques
●
●
Institutionalization of
data sources
●
●
Quality control
●
●
●
Training
Legacy data
●
●
●
●
●
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
Broader Implications



These issues apply to any effort to create
large, federated data sets
NSF and other federal agencies will be pushing
hard on these issues
Cooperation with private sector will be critical
–
–
Both have common needs
That’s likely to be where the technology will come
from
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN
For Further Information

Science of Collaboratories Project
–

www.scienceofcollaboratories.org
Collaboratory for Research on Electronic Work
(CREW)
–
www.crew.umich.edu
SCHOOL OF INFORMATION
UNIVERSITY OF MICHIGAN