Transcript 20030409-Cyberinfrastructure-Blatecky

Cyberinfrastructure Implications for the Future of Research
Alan Blatecky
ANIR
National Science Foundation
Spring 2003 Internet2 Meeting
Driving Factor 1:
Technology Pull
• Continuing exponential advances in sensor, computer, storage
and network capabilities will occur
• Sensor networks will create experimental facilities
• PetaByte and ExaByte databases will become feasible
• Increase in numerical and computer modeling capabilities
broaden the base of science disciplines
• Increase in network speeds makes it feasible to connect
distributed resources as never before
Spring 2003 Internet2 Meeting
Driving Factor 2:
Science Problem Push
• New classes of scientific problems are enabled from
technologies development
– High energy physicists will harness tens of thousands of CPUs in a
worldwide data grid
– On-line digital sky survey requires mechanisms for data federation
and effective navigation
– Advances in medical imaging and technologies enable
collaboration across disciplines and scale
• Coupling of expertise, collaboration, and disciplines
encourage the development of new science and research
Spring 2003 Internet2 Meeting
Future Science Scenarios
• Virtual observatory – graduate student in Brazil is struck by a strange spectrum of an object she
had read in a paper, and uses existing data to search for matching
characteristics and is able come up with interesting results and
observations
• Digital medical imaging and personalize medicine
– advances in imaging technologies are utilized for routine screening, early
assessment and the design of prevention therapies
• Disaster Response
– hurricane predicted to hit New Orleans at the same time a terrorist anthrax
threat in Houston has been confirmed; FEMA uses national computing,
database, sensor resources and so forth to determine appropriate response
strategies
Spring 2003 Internet2 Meeting
Scenarios continued
• Distributed collaborative designs for safer
buildings
– testing complex multi-component structures and associated reactions with
sensor data to design safer structures
• Large-Scale international science
– A $5B fusion reactor can be operated efficiently as an international
facility; experiments can be designed, configured and monitored by a team
in a virtual control room.
• National Airspace simulation
– The entire national airspace can be modeled and simulated including
weather, real-time data, terrain data, radar and sensor data
Spring 2003 Internet2 Meeting
Observations
• Recent changes in the nature of science are creating an urgent
requirement for a new class of distributed science infrastructure
• Infrastructure includes an array of grid and middleware
technologies
• E-science has some unique requirements that will not be
addressed by industry
• Current Grid/Middleware technology and software development
is being done via a patchwork of diverse, short-term projects and
programs
Spring 2003 Internet2 Meeting
The NSF Objective for
Cyberinfrastructure
• provide an integrated high end system of
hardware, software, and services that ...
• enables scientists and engineers to work on
advanced research and education problems that
would not otherwise be solvable
Spring 2003 Internet2 Meeting
Cyberinfrastructure
Spring 2003 Internet2 Meeting
Cyberinfrastructure:
an Ecological Model
Expertise
Scientific Disciplines
Research
Education
Technical support
Software Development
Grid Operations
Organizations
Universities, Schools
Government Labs
Government Agencies
Research Centers
Libraries, Museums
Results
Network Resources
Collaboration
Shared Resources
Virtual Organizations
Production Backbones
Experimental Networks
Research Networks
Middleware services
Grid capabilities
Computational Resources
Supercomputers
Grid Tools
Visualization
Computing Clusters
Specialized Architectures
Data & Knowledge Resources
Databases
Storage
Collections
Mining & Navigation
Tools
Spring 2003 Internet2 Meeting
New Science
New Applications
Grand Challenges
Competitiveness
Societal Impact
Scientific Instruments
Telescopes
Colliders
Shake Tables
Microscopes
Sensor Arrays
Some Issues
• A recognition that long-term, sustained, and persistent efforts to
develop and support grid and middleware must be established
and encouraged as soon as possible
• Inadequate funding of Grid and middleware in general;
inadequate funding of pipelines to produce expertise for future
• A recognition that international coordination, cooperation and
collaboration is essential
• Significant federal research support and development is required
immediately, including a long term development plan
Spring 2003 Internet2 Meeting
Areas requiring immediate
attention and support
• Robust Software and Middleware
– “production” software engineering; including usability
– hardened infrastructure
– scaling issues, global deployment
• Grid Management and Operations
– heterogeneous, multi-domain coordination of resources
– authentication, authorization, monitoring, etc
• Grid interoperability
– systemic programs and interoperability testbeds
– development of consensus/standards
– international coordination
Spring 2003 Internet2 Meeting
Necessary Long-term Investments
•
•
•
•
Fundamental CS&E research
Operational cyberinfrastructure
Shared cybertools (software)
Domain-specific cybertools
To support research and education in all science and
engineering fields.
Spring 2003 Internet2 Meeting
Cyberinfrastructure Drivers
• Extensible Terascale Facility (ETF)
• Partnerships for Advanced Computational Infrastructure
(PACI)
• Grid Physics Network (GriPhyN)/ iVGDL
• Network for Earthquake Engineering Simulation (NEES)
Grid project
• National Virtual Observatory (NVO)
• National Ecological Observatory Network (NEON)
• National Science Digital Library (NSDL)
Spring 2003 Internet2 Meeting
Technical Challenges
• More players, more heterogeneity, design must scale
• System must be production-ready – SW must be bullet-proof,
useful, usable
• Most interesting programming models (real-time, on-demand,
adaptive, etc.) still require considerable research
• Programming environments, tools must be developed and
usable
– How do we debug a Grid program?
Spring 2003 Internet2 Meeting
Logistical, Legal Challenges
• Who maintains the Software? Who fixes the bugs? Who
documents the code? Who answers the phone? Who wears
the pager?
• How do we do accounting over multiple administrative
domains? How do we allocate resources over multiple
sites?
• How do we deal with varying institutional IP policies, open
source policies, licensing policies, etc.
• What authority and responsibility will Virtual
Organizations have?
Spring 2003 Internet2 Meeting
Ideological Challenges
• What organizational framework promotes development of
stable, persistent infrastructure?
• How to integrate different institutional approaches and
cultures for administration of resources, operations,
Software development and deployment, etc.
• How to develop metrics and incentives for meaningful
cooperation and coordination
• What is shared/private, free/charged for,
centralized/distributed, etc.
• How do you share resources across national boundaries?
Spring 2003 Internet2 Meeting
International Challenges
• Who makes decisions? Who enforces decisions?
– Do we need a “Cyberinfrastructure UN”?
• How do we do global allocations?
• What mechanisms should be used to select/support
applications?
• How do we ensure stability and interoperability?
• How do we ensure that the sharing of resources
does not compromise security?
Spring 2003 Internet2 Meeting
end
Spring 2003 Internet2 Meeting