Organization of an Advanced Cyberinfrastructure Program ACP

1
Organization of an Advanced Cyberinfrastructure
Program (ACP)

• by
• David G. Messerschmitt
• Member, NSF Blue Ribbon Panel on
  Cyberinfrastructure

2
Methodologies of investigation
Experimentation
Instruments, sensors and data analysis
Theory
Modeling and visualization
3
What is digital science?

• Vast collection and repositories of data
• Modeling, computing, visualization
• Distributed instrumentation
• Scholarly publishing
• Distributed software applications
• Collaboration

4
Horizontal layering dimension
Sensors, instrumentation
Scientists
Conduct of science and engineering research
Distributed applications
Cyberinfrastructure shared resources and services
5
Technology transfer dimension
R
Conduct of science and engineering research
Distributed applications
Cyberinfrastructure shared resources and
services
Research Development Operations

support
6
Opportunities

• Rethink and reorganize the processes of research
  for effectiveness and efficiency by exploiting IT
• Empower researchers to benefit from IT without
  becoming IT experts
• Enable the emergence of new disciplines and
  multi-disciplinary and cross-disciplinary efforts
• Capture, preserve, and publish outcomes for
  future critique and extension
• Stimulate the advance of IT and the Nations
  economy

7
Rethink and reorganize the processes of research

• Significant value rarely derived from automation
  alone
• Significance of place and institution can be
  reduced level the playing field
• Role for domain scientists and engineers with
  assistance of computer scientists and social
  scientists
• Requires visionary leadership in each Directorate

8
Empower researchers to benefit from IT without
becoming IT experts

• Testimony indicates we are moving in the wrong
  direction
• Research, infrastructure, tools, design, etc.
• Not just technicalpeople, organizations
• Support staff and staff continuity

9
Enable the emergence of new disciplines and
multi-disciplinary and cross-disciplinary efforts

• Infrastructure enables application
  interoperability
• Extract pan-scientific commonality and standards
• Tension with individual initiative
• Tension with desire to research thoroughly before
  standardizing
• ? Spiral approach

10
Capture and communicate outcomes for future
critique and extension

• Preservation a major new responsibility
• Data will almost always be intermediated by
  software
• Preservation of software behavior a major
  technical challenge

11
Stimulate the advance of IT and the Nations
economy

• Buy and integrate (rather than make) where
  appropriate
• Coordinate with and influence direction of
  industry
• Commercialize and shift funding to the frontier
  where possible
• Open source (w/ governance and support) where not
  commercializable

12
Centralization

• Operations
• Processing (modeling and data analysis)
  supplementing local and grid resources
• Software-intermediated data (preservation and
  access) for incorporation into distributed
  applications
• Development Standards, software infrastructure
  and tools, software integration
• Support software distributions and operations

13
Within NSF
Directorates
CISE
Applications
Commonalities
Choose technologies
Metholology and processes of research
Advance IT itself
14
ACP Program
ACP Office and leadership
15
Leadership
Leader of ACP Office
Budget Congress
-----------------Coordinate----------------------
Directorates Agencies International

• Functional peer with the assistant directors of
• the NSF directorates
• To attract high caliber individual
• Avoid conflicts of interest

16
Outside standards and software
Systems engineering and integration
Processing infrastructure and tools
Storage infrastructure, standards, and tools
Communications middleware
Geophysics infrastructure
Astronomy infrastructure
Etc.

17
Co-location acceptable
Systems engineering and integration
Processing infrastructure and tools
Storage infrastructure, Standards, and tools
Communications middleware
Geophysics infrastructure
Astronomy infrastructure
Etc.

18
Operations and support
Processing operations
Storage operations
Communications operations
Processing support
Storage support
Communications support
Integrative support
Users
Disciplinary centers
19
Operations data centers
Physical data stewardship
Geophysics infrastructure
Astronomy infrastructure
Etc.
Disciplinary content curation
20
Conclusions

• A major opportunity, and also major challenges
• Costs of not moving aggressively fragmentation,
  inflexibility, duplication of effort, lowered
  efficacy and effectiveness
• About 1B/year added budget is mostly human
  resources (development, operations, support)
• NSF should take the lead, but needs to work with
  other agencies and international bodies