Cyberinfrastructure

November 16, 2006

National Science FoundationDirectorate for Engineering

Abhijit DeshmukhPD, Manufacturing Enterprise Systems/CMMI Office of CyberinfrastructureENG Cyberinfrastructure Working Group

Complex, multi-scale, multidisciplinary S&E research challenges

Complex, multi-scale, multidisciplinary S&E research challenges

Advances in components of CI-

systems for S&E R&E

Advances in components of CI-

systems for S&E R&E

30+ disciplinary workshops on CI vision & impact

30+ disciplinary workshops on CI vision & impact

NSF internal working groups

NSF internal working groups

NSB & Community

Input

NSB & Community

Input

CI Council, Directorate/Office CI Activities, OCI,

ACCI

CI Council, Directorate/Office CI Activities, OCI,

ACCI

NSF Cyberinfrastructure Vision

All directorates and offices support cyberinfrastructure.

Science-driven partnerships between creation, provisioning and use of CI

Supports integrated research and education and broadened access and participation.

High Performance Computing

High Performance Computing

Data, Data Analysis &

Visualization

Data, Data Analysis &

Visualization

Virtual Organizations

Virtual Organizations

Learning & Workforce

Development

Learning & Workforce

Development

Vision Framework

Draft available at www.nsf.gov/oci/

High High PerformanPerforman

ce ce ComputingComputing

High High PerformanPerforman

ce ce ComputingComputing

Life

Satellite tobacco mosaic virus, P. Freddolino et al.

Aldehyde dehydrogenase, T. Wymore and S. Brown

Matter

I. Shipsey

The EnvironmentSociety

John Q Public

S.-Y. Kim, M. Lodge, C. Taber.

increasingly important tool for understanding

Track 1: One solicitation funded

over 4 years: $200M acquisition +

additional O&M cost.Track 2: Four

solicitations over 4 years: $30M/yr

acquisition + additional O&M cost. First track 1

approved 8-07

Data, Data Data, Data Analysis & Analysis & VisualizatiVisualizati

onon

Data, Data Data, Data Analysis & Analysis & VisualizatiVisualizati

onon

• Challenges: increased scale, heterogeneity, and re-use value of digital scientific information and data. Inadequate digital preservation strategy of long-lived data.

• Taking initial steps to catalyze the development of a federated, global system of science and engineering data collections that is open, extensible, evolvable, (and appropriately curated and long-lived.)

• Complemented by a new generation of tools and services to facilitate data mining, integration, analysis, visualization essential to transforming data into knowledge.

• NSF Leadership for OSTP/Interagency Working Group on Digital Data

• Distributed virtual organizations are based upon CI that provides flexible, secure, coordinated resource sharing among dynamic collections of individuals, institutions, and resources.

• Resources and services include HPC, data/information management, sensor-nets/observatories, linked through global networking and middleware, and accessed by people through web portals and workflow environments.

• Challenges being address include tools for more rapid building and ease of use, interoperability/middleware, high performance, end-to-end networking, and dynamic reconfiguration, social issues, assessment of impact, and economic and technical sustainability.

Virtual Virtual OrganizatioOrganizatio

nsns

Virtual Virtual OrganizatioOrganizatio

nsns

NanoHubNEES

ATLAS

NVO

LEAD

iVDgL

Learning & Learning & Workforce Workforce DevelopmDevelopm

entent

Learning & Learning & Workforce Workforce DevelopmDevelopm

entent

•Learning supported by CI. (cyber-enabled learning).

•Workforce development to create and use CI for S&E research and education.

•Broadened participation: Exploit the new opportunities that cyberinfrastructure brings for … people who, because of physical capabilities, location, or history, have been excluded from the frontiers of scientific and engineering research and education.

•Explore CI support for integrated research and education.

CyberBridgesMARIACHI

EPIC

BIOINFORMATICS CI INSTITUTE

Duality of CI CI for Big Science

Flops: Mega, Giga, Tera, Peta, … Large data sets Fire and forget – batch processing Size, latency, resource discovery are NOT critical

CI for Pervasive Use Artificial nervous system for engineered systems Facilitates creation of adaptable/ reconfigurable/ evolvable

systems Control and coordination: computational and non-

computational resources Size, latency, bandwidth, dynamic composability, real time

constraints are VERY important

CI and ENG

ENG believes that cyberinfrastructure will play an essential role in future engineering-related research.

ENG’s needs in this realm often dovetail with other directorates; however, ENG also has unique requirements and leadership capabilities: Sensor networks and real-time issues Design and control of complex systems Multi-scale phenomena

ENG can contribute significantly to the creation of this infrastructure Novel processors, data storage and networking Design, management and control of CI

ENG Investments in CI (FY 04-06 by Category)

Legend

HPC: High-performance ComputingData: Data, Data Analysis and VisualizationVO: Virtual OrganizationsWorkforce: Learning and Workforce

DCI: Development and Deployment of CIRCI: Research Enabling CIUCI: Use of CI

The final ENG Totals column is truncated to preserve scale, and totals $117,317,000.

ENG CI Priorities (FY 2007 – 2012)

Engineering Research Frontiers Enabled by CI

(EFRI: ARES-CI)

Engineering Gateways/ Virtual Organizations

(Seed Grants)

Multi-scale, Multi-phenomena Modeling (SBES)

Research Enabling Next Generation CI Capabilities

(PetaApps)

CI Education and ENG Education using CI (CIEG)

Autonomously Reconfigurable Engineered Systems Enabled by CI (ARES-CI)

From Complexity to Reconfigurability Complexity arises from the need to be robust in presence of

anticipated faults Complex systems are robust to known uncertainty – yet fragile to

unknown events Reconfigurable or topologically modifiable systems enable

robustness to unknown failures

Core Unanswered Questions What are the fundamental principles underlying design and control

of reconfigurable systems? How much reconfigurability is enough? What/when to change/reconfigure? Continuum of adaptability, reconfigurability and evolvability

ENG

NSF 06-596

Engineering Gateway Seed Grants

Virtual organizations (VOs): Communities of researchers and educators linked by CI resources Can play an important role in promoting collaboration

Early NSF experience with gateways has been very positive nanoHUB.org for nanotechnology researchers NEES for earthquake engineering researchers

NSF Engineering is considering seed grants to assist communities form VOs

Anticipated Activity

ENG

Overarching Framework for Multiscale Modeling: atomistic micro meso macro

Key Issue: interfaces/exchanges between models at different length and time scales

Questions:

• What information needs to be transferred from one model segment to another?

• What are the correct and most effective ways to achieve such transfer of information?

• What physical principles must be satisfied during the transfer of information or simulation results?

Need a set of consistent logical, mathematical, and physical rules to govern information transfer across the interfaces

Simulation-based Engineering Science

CI will help catalyze a transformation to high-fidelity, simulation-based engineering science.

Simultaneous advances of the models, methods and algorithms that underpin the components are crucial for realizing the potential of CI.

Example: Simulation-based planning for vascular bypass surgery. From left: (1) MR image data, (2) preoperative geometric solid model, (3) operative plan, (4) computed blood flow velocity in aorta and proximal end of bypass, and (5) postoperative image data used to validate predictions. (UTA and Stanford)

WTEC study planned in FY 07

Accelerating Discovery in Science and Engineering Through Petascale Simulations and Analysis (PetaApps)

Anticipated program size: $15M Expected award amounts: up to $1,000,000 Potential proposal topics:

Enhancing algorithmic scalability exploiting multi-threaded, highly parallel, hierarchical architectures

Improving and creating data sampling, analysis and clustering algorithms for large data sets

Developing innovative modeling, simulation or optimization algorithms suitable for petascale systems

Innovative computational techniques that were previously not viable due to hardware capability

Anticipated Activity

OCI, ENG, MPS

CI Experiences for Graduate Students

“Boot camp” for cyberinfrastructure Goal: Train engineering PhD students in CI tools and

techniques

Pilot program in Summer 2007 Summer residency at the San Diego Supercomputer

Center Supplements to existing MES, SEE, OR grants Supplement request deadline: December 1, 2006 Anticipate expanding to other facilities and other

programs in the futureENG

Dear Colleague Letter NSF 06-044

Cyberinfrastructure Training, Education (CI-TEAM)

• Goals: Develop a diverse cyberinfrastructure workforce Foster inclusion in cyberinfrastructure activities of diverse groups

• FY06 program funds ~ $10 M for two types of awards: Demonstration Projects ≤ $250,000 Implementation Projects ≤ $1,000,000

• Demonstration Project: Exploratory with the potential to serve as pathfinder for larger-scale implementation activities in the future

• Implementation Project: Expected to deliver sustainable learning and workforce development activities that complement ongoing NSF investment in cyberinfrastructure

• Multidisciplinary teams, significant impact from partnerships

• Leveraged cyberinfrastructure, replicable and (potentially) scalable

Anticipated in FY 07 (past solicitation NSF 06-548)

OCI

Other Opportunities

Two target dates each year:

2nd Thursday in February & August

Unsolicited proposals for the development and/or demonstration of CI services and resources or for CI education, outreach and training activities that fall outside the scope of other programs at NSF or elsewhere.

Strategic Technologies For CI (STCI)

OCI

Standing Program

Software Development For CI (SDCI)

Full Proposal Deadline: January 22, 2007 Program size: 10 to 20 awards, $14M total funding Award amounts:

$50,000 - $1,000,000/year, 2-3 years Focus areas for FY07:

High Performance Computing (HPC) environments Digital data acquisition, discovery, access, analysis, and

preservation Middleware capabilities and services to support distributed

resource sharing and virtual organizations

Program Solicitation NSF 07-503

OCI

National Digital Data Framework Concept

Digital repositories launched with explicit goal to achieve long term sustainability

Expertise in: Cyberinfrastructure, library and archival sciences, data

science, computer and information science, social and behavioral sciences, economics, domain sciences

Combination of awards designed to test: Sustainability models Economies of scale New partnerships across sectors

Anticipated Activity

OCI

ENG Community Input

ENG Ad Com Subcommittee on Cyberinfrastructure: “A Process-Oriented Approach to Engineering Cyberinfrastructure” Assessment: CI user requirements, resource tracking, infrastructure usability

metrics Coordination: with OCI, other directorates, other agencies Planning: ENG priorities Building the “Innovation Loop”: ENG CI research challenges, synergies

between CI research, development and deployment ENG sponsored community-wide workshops and reports on CI strategies

ENG communities need unique engineering gateways that focus on different communities

Blue Ribbon Panel recommends a significant investment in multi-scale, multi-phenomenon modeling across the engineering disciplines

ENG should invest in creation of focused “facilities" (e.g., hazards, sensors, or environmental observatories) that will enable frontier research in different disciplines

Thank you