NEES Networking Needs The NEES MRE: Where the Infrastructure Community Meets the Cyberinfrastructure Community

1
NEES Networking NeedsThe NEES MRE Where the
Infrastructure Community Meets the
Cyberinfrastructure Community

• Kyran (Kim) Mish, Director
• Center for Computational Engineering
• Lawrence Livermore National Laboratory

2
So Exactly What is NEES?

• Network for Earthquake Engineering Simulation
• NEES is a distributed array of experimental
  sites, grid-based data repositories, tool
  archives, and computational resources, all
  seamlessly linked (hopefully!)
• NEES has four components, with three now funded
• The consortium, which will run NEES after 2004
• The consortium development (CD) builds the
  consortium
• The experimental sites, which provide data and
  content
• The systems integration (SI) effort, termed
  NEESgrid
• Network drivers include telepresence, curated
  repositories, scalable HPC, experimental-numerical
  coupling, short- and long-term QoS issues

3
Example NEES Experimental Site

• Geotechnical Centrifuge at UC Davis

4
NEES Network Stakeholders

• Experimental Facilities
• Shake tables, centrifuges, wave tanks, field
  sites
• Resource providers
• Computers, software, storage, networks
• End users
• Researchers, practicing engineers, students,
• Operational facilities
• NCSA/NEESgrid ? NEES Consortium in 2004

5
Typical NEES Cyberinfrastructure
6
Infrastructure vs. Cyberinfrastructure

• Characteristics of Infrastructure Culture
• Risk averse, which leads to slow technology
  adoption
• Code-based practice to defend against litigation
• Follow community wants/needs whenever possible
• Goal is highest reliability, e.g., MTBF
• Characteristics of Cyberinfrastructure Culture
• High-risk, innovate or die trying approach to
  technology
• Best-practices approach leaves legal issues
  dangling
• Develop technology, then look for a market
• Goal is highest performance, e.g., TFLOPS

7
Typical NEES Infrastructure

• Infrastructure community builds ubiquitous
  networks
• Robust, reliable, redundant, extensible over time
• Generally, these networks degenerate gracefully
  with load
• High-value packets are seldom lost, thankfully

8
Consider Science and Engineering

• Science is a process whose desired outcome is
  scientific truth
• Open sharing of data in community of science
• Metric is evidence of a creative mind
• Engineering is a profession whose desired outcome
  is technology
• Information may be proprietary, IP dominates
• Metric is financial or market-driven (share)
• NEES MRE must respect these differences

9
Questions, Answers, and Comments

• Words of Wisdom from Bill Lennon, LLNL

People put up with networks only because they
are a necessary evil