Collaboration across the Pacific

1
Collaboration across the Pacific

• John Silvester
• PI, NSF-OCI-IRNC TransLight/PaificWave
• Professor of Electrical Engineering, USC
• Board Member, CENIC
• silvester_at_usc.edu

CENIC HPR/XD Workshop, CalIT2, September 16, 2008
2
Asia Pacific Networks
3
Pacific Wave www.pacificwave.net

• Pacific Wave is an exchange supporting the
  interconnection of research and education
  networks
• Allows US RE networks to peer with foreign
  networks (pass traffic to each other) over
  international links
• Pacific Wave has 3 points of presence Los
  Angeles, Sunnyvale, and Seattle
• Mutiple 10G links between the PoPs
• Research connection (over C-Wave) to Starlight
• Operated by CENIC and Pacific Northwest GigaPop

4
RE Networks and Pacific Wave(Layer 3 and Layer
2 Interconnects)
Canadian Net
Australasian Nets (mostly)
CANET4
AARNET GEMNET GLORIAD IEEAF KREONET MIMOS
(Malaysia) QATAR SINET SINGAREN TANET/TWAREN TRA
NSPAC2
US Nets
Pacific Wave
INTERNET2 CALREN (HPR) DREN ESNET NASA NLR NREN P
NWGP
CLARA CUDI
Latin America Nets
5
Pacific Wave
6
TransLight/PacificWave

• TransLight/PacificWave (TLPW) is a US-NSF-OCI
  IRNC project.
• Primary Objectives are
• Assist in the termination (equipment, backhaul)
  of the Australian Research Network links from
  Sydney to the US West Coast (pays for the US
  portion of the links)
• Provide funding to allow the links to stop off in
  Hawaii (on Oahu and the Big Island)
• Provide funding to upgrade connectivity to the
  (international) observatories on Mauna Kea
• Provide partial support for the operation of
  Pacific Wave exchange facility
• Provide funding for upgrading the Pacific Wave
  Exchange to enable newer technologies including
  high bandwidth switched connections for demanding
  applications
• Provide funding for outreach and support of the
  user community (science and engineering
  researchers), especially in Australia and New
  Zealand (collaboration)

7
AARnet SX-Transport
8
Application Development and Support

• Organized several application oriented workshops
  for different discipline areas by attaching a
  networking workshop to an pre-existing scientific
  meeting.
• Add-on workshop for GRIDAsia 07
• Participation in KAREN (NZ) workshop to
  demonstrate/explore opportunities of high speed
  networks
• AARNet/APAC/TLPW Joint Workshop after APAC
  meeting (October 2007) Driving e-Research
  Collaborations Across the Pacific

9
DeRCAP www.apac.edu.au/apac07/dercap/

• Held in conjunction with the APAC meeting in
  Perth (WA)
• Identified several disciplines and key Australian
  scientists in each area
• Asked them to identify current or potential US
  partners
• Invited US partners to participate in workshop
• Workshop focused on challenges faced by research
  collaborations to give input to the
  infrastructure community as to what is needed

10
Areas Covered Marine Systems

• John Orcutt, Director, UCSD Center for Earth
  Observations and Applications
• Craig Johnson. University of Tasmania

11
Earth Systems and Climate Science

• Tim Pugh, Bureau of Meteorology
• Peter Fox, OpenDAP
• Nathan Bindoff, Director, TPAC, University of
  Tasmania

12
HEP

• Shawn McKee, University of Michigan
• Glenn Maloney, University of Melbourne

13
Astronomy

• Peter Quinn, Premier's Fellow, The University of
  Western Australia
• Colin Lonsdale, MIT, Haystack

Telescopes are usually In remote
places! Population density 0.003 / sq km
14
Remote Instruments

• Rick McMullen, Director, Knowledge Acquisition
  and Projection Lab, Pervasive Technology Labs,
  Indiana University
• Peter Turner, University of Sydney

15
Geosciences and Biosciences

• Geosciences
• Dion Weatherley, Earth Systems Science
  Computational Centre, University of Queensland
• Rob Woodcock, Exploration and Mining, CSIRO
• John Rundle, Centre for Computational Science and
  Engineering, UC Davis
• Bio Sciences
• Matt Bellgard, Director, Centre for Comparative
  Genomics, Murdoch University

16
Generic Needs

• Shunting large amounts of data round
• Big science
• Large data streams for sensor networks (climate
  change, disaster and natural hazard warning
  monitoring systems, emerging infections and
  biosecurity)
• Widespread access for sensor nets, remote
  facilities, users, and applications
• Immersive remote collaboration the higher the
  resolution the better
• Simple to use tools enabling the infrastructure
  (middleware)
• Disaster recovery collaboration

17
Issues Identified Network and Bandwidth

• Challenge of BW to remote locations (astronomy,
  SKA, environmental sensing)
• V High BW challenges down the road (SKA, others)
• Better problem identification and isolation
  (measurement and monitoring infrastructure)
• Discipline investment allows for more effective
  usage (cf HEP)
• Shared infrastructure has great value.
• Commercial players. Google? Etc.

18
Issues Identified Support Services

• Network support during SW development process
  (for remote instruments)
• How to transfer expertise in architecture from
  one domain to another
• Better problem identification and isolation
  (measurement and monitoring) 
• Need to have the technical expertise available
• Need people that can bridge network, computing
  and discipline
• Central help desk single point of contact
• Regular operations meetings which includes
  applications people

19
Issues Identified Middleware I

• Lots of middleware issues AAA
• Different levels of challenges policy,
  architecture, standards, middleware,
  infrastructure
• Data preservation and backup 
• Missing levels in the solution stack. This could
  allow for development of commonality. What is the
  boundary between common infrastructure and
  applications?

20
Issues Identified Middleware II

• Still in the stage of competition in the
  middleware and tools - point solutions
• Shared infrastructure has great value. Federate
  infrastructure providers to leverage
  capabilities?
• (Discipline specific) data architectures
• Importance of monitoring tools
• Dashboard for an application domain, science
  gateways

21
Summary

• Consensus was that todays networks are generally
  adequate for todays needs (except for access to
  remote locations) BUT some large data
  generators/consumers coming on-line in the next
  few years
• Need to be make the infrastructure more easily
  usable and accessible to researchers (middleware)
• Need for multi-disciplinary support teams
• Need for outreach to researchers what is
  possible?

22
Asia Pacific Networks