National Workshop on Cyberinfrastructure

1
National Workshop on Cyberinfrastructure
OneTenn for Tennessee
Douglas E. Hurley University of Memphis

• May 12, 2006
• Nashville, TN

2
Topics

• What is Cyberinfrastructure
• The process to develop OneTenn
• OneTenn what it is/why is it important
• Focus on - Real research, Real value
• What would OneTenn look like

3
First - What is Cyberinfrastructure ?

• new research environments that combine
  computation, information management, networking,
  and intelligent sensing systems
• leading to more broadly trained scientists,
  engineers, and technical personnel with blended
  expertise in interdisciplinary science or
  engineering, mathematical, and computational
  modeling, numerical methods, and visualization
  and who possess the sociotechnical understanding
  about working in new grid or collaboratory
  organizations.
• Excerpts from the NSF Cyberinfrastructure Report

4
What should Cyberinfrastructure do?

• The Challenge
• The size and data intensity of the problems to
  be solved outstrips the capacity of private or
  individual local resources.
• Access to resources
• processing power, bandwidth, data storage,
  visualization tools
• move, store, process, and share large data sets
• Enable collaboration across boundaries
• different disciplines and institutions.
• expertise is rarely co-located in a single place
  or organization.
• research teams work across physical (and
  political) boundaries
• Incorporate innovation rapidly
• New science brings new methods and techniques
• Requires rapid access and implementation
• Open standards and technical interoperability
• Overcome political as well as technical barriers

5
Cyberinfrastructure as defined in OneTenn

• the common and persistent base of computational
  and communication resources hardware, software,
  and people shared by a community in order to
  facilitate well coordinated interactivity in some
  area of human endeavor that is highly compute and
  data intensive.
• OneTenn Cyberinfrastructure report

6
The process that led to OneTenn

• Understanding cyberinfrastructure
• Awareness and information phase
• Survey of statewide computational and storage
  resources
• Hard to gather this information
• Extended discussion and debate
• A computation grid and storage services, or a
  network or both
• Attempted to engage research leadership
• Partial success
• Solicited "meritorious applications
• Took awhile

7
OneTenn Vision

• OneTenn
• builds upon Internet-based principles for
  interoperability
• brings together 2 essential cyberinfrastructure
  resources
• very high speed networking and very large data
  storage and staging services
• provides easy to use shared access to critical
  computation resources
• By deploying to every four year and two year
  college and university, OneTenn will enable
  advanced research and education across the state
  to work together as one community, as One
  Tennessee.

8
No Field of Dreams

• Focus on applications and services
• Not on big infrastructure
• Its important to have applications and services
  running quickly after infrastructure deployment

9
OneTenn is different from other
cyberinfrastructure initiatives

• Balanced approach Deploy integrated optical
  networking with highly scalable network storage,
  to enable access to computation assets
• Why is storage important?
• Storage is fundamental to the most powerful
  network apps e-mail and the Web
• Data intensive applications require even more
  working storage
• Whats different about IBP for storage?
• Generic, best effort storage service (read
  write)
• Designed for interoperability and
  scalability/end-to-end design
• Middleware to use it already available
• Access to (not deploying) computation
• Unlike networks and storage, computation is not a
  generic resource 
• Different kinds of applications require different
  kinds of systems
• Systems don't easily interoperate

10
OneTenn is designed to expand interoperability
11
Real research, Real value

• Tennessee Data Intensive Collaborations
• Particle Physics with Fermi (Vanderbilt/UTK)
• Computational Analysis of Gene Regulatory
  Networks (UTK, UTHSC, ORNL)
• Distributed Data Mining for Genomics (UTK/UTHSC)
• Distributed Visualization for Medical Imaging
  (Vanderbilt/UTK)
• Optical Networking for Remote Instrument Control
  (UoM, Vanderbilt)
• Neural Networks Analysis of Microarray Data
  (ETSU)
• Other research applications Supernova
  simulation, Fusion energy simulation,
  Environmental modeling
• Educational/other Applications Video-on-demand,
  Web-cast capture and replay, video transcoding,
  content distribution, telemedicine

12
The OneTenn infrastructure upon completion of the
five year plan

• The OneTenn network with storage depots directly
  attached.
• 4 year sites will connect at 10Gbps 2 year sites
  will connect at 2.4 Gbps.
• Total storage available in aggregate 1PB.
  Approximately 550 TB's distributed among twelve
  4-year schools, 140 TB's at 13 2-year campuses,
  and 340 TB's at the Tennessee Technology Centers
  and early adopter P-12 sites.
• While the Tech Centers and participating P-12
  sites, are not connected to the backbone, and
  hence not shown on the map, the depots at their
  sites will facilitate their participation and
  increase their ability to benefit from OneTenn,
  especially in the area of multimedia content and
  video distribution.

13
Q/A

• Doug Hurley
• VP, Information Technology/CIO and Chairman,
  FedEx Institute of Technology
• University of Memphis
• [phone]
• dhurley_at_memphis.edu