TransPAC HPIIS Performance Review

1
TransPACHPIIS Performance Review

• Michael A. McRobbieVice President for
  Information Technology and CIO
• Indiana University
• October 25, 2000

2
Overview

• HPIIS, TransPAC and APAN
• The TransPAC Cooperative Agreement
• Progress toward objectives
• Science enabled by TransPAC

3
Goals of the HPIIS program

• US NSF Program for High Performance International
  Internet Services (HPIIS), solicitation
  NSF97-106.
• The goal is to provide the basis for the
  development of next generation applications
  supporting international collaborations and
  enabling interaction with global information and
  experimental resources.
• HPIIS connections link international
  high-performance production networks for
  international scientific, research and
  educational collaborations. Additionally, HPIIS
  networks enable global testbeds for new
  networking technologies.
• e.g. multicast, IPv6, QoS and web cache

4
Historical Overview

• April 1997 HPIIS solicitation
• August 1997 IUs TransPAC response
• February 1998 TransPAC Proposal revised
• August 1998 Cooperative Agreement signed
• August 1998 TransPAC operational

5
Asia-Pacific Advanced Network (APAN)

• 14 member-nations cooperating to build a high
  performance research and education network across
  the AP region
• 5 primary members
• Australia, Japan, Korea, Singapore, USA
• 2 associate members
• China, Malaysia
• 7 affiliates, liaison and other members
• Hong Kong, Indonesia, Thailand, Philippines,
  Canada, Europe (Dante), Consultative Group on
  International Agricultural Research (CGIAR)

6
TransPAC Connecting APAN to the US
7
STAR TAP Logical Map
8
TransPAC Objectives

• Provide leading edge connectivity between vBNS
  and APAN
• High throughput production network for research
  and education
• Testbed for new protocols, network services and
  applications

9
TransPAC Consortium

• Indiana University (lead US organization)
• APAN (APAN-JP lead organization)
• National Science Foundation HPIIS program
• Japan Science and Technology Agency (JST)
• Kokusai Denshin Denwa, Co. Ltd (KDD)
• ATT
• Korea Telecom
• STAR TAP
• NCSA and NPACI

10
TransPAC Cooperative Agreement
11
Cooperative Agreement

• TransPAC operates with funding from
• The NSF
• the Japan Science and Technology Agency
• additional contributions from Indiana University
• Operation is controlled by NSF CA ANI-9730201

12
Terms of the Cooperative Agreement

• Primarily an agreement for services
• 5 year term, renewable annually
• Specific deliverables based on HPIIS program
  objectives
• General notion of cooperating with other HPIIS
  funded projects
• Review at the beginning of the third year
• Changes in RE networks
• Impact on science

13
Cooperative Agreement Deliverables

• Provide and manage a direct, high-speed network
  between the vBNS and APAN
• Allow access only to HPIIS approved institutions
• Monitor the performance and use of the TransPAC
  connection
• Cooperate with NLANR to develop and test new
  Internet protocols
• Provide publicly-accessible information about
  TransPAC and the projects it enables
• (a) In concert with NLANR provide consultative
  user services and (b) front-end tools for direct
  access to differentiated network services

14
1. Provide and manage a direct, high-speed
network between the vBNS and APAN

• Network infrastructure
• Chicago-Tokyo link provided through annually
  renewable contract with ATT and KDD
• Peering at STAR TAP and the APAN Tokyo Exchange
  Point
• Infrastructure jointly funded by the NSF, JST
• Network engineering
• Provided by IU, STAR TAP, and KDD staff
• Funded by NSF, IU and KDD (in Tokyo)
• Network operation and information centers
• US side Indiana University (Global NOC)
• APAN side APAN/KDD Tokyo XP
• User support
• US side NSF and IU funded (IU and NLANR DAST
  staff)
• APAN side provided by the APAN community

15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Highlights

• Network and support staff jointly funded by NSF,
  JST and IU
• TransPAC was up and on-line before the
  cooperative agreement was signed
• Partnership with JST, ATT and KDD have led to
  continuous improvements in service levels without
  additional NSF funding
• Synergies within the HPIIS program and with
  Abilene/Internet2 have led to the Global NOC, run
  by Indiana University

19
TransPAC No-Cost Bandwidth Improvements
OC-3 southern route (Q2/2001)
155 Mbps upgrade (10/00)
ATT consolidation upgrade (11/99)
73 Mbps JST upgrade (5/99)
Initial 35Mbps
20
(No Transcript)
21
Leveraging the HPIIS TransPAC CA

• First circuit
• NSF 2M annually for 5 years (US half circuit)
• JST 3M annually (est.) for 5 years (JP half
  circuit)
• Second circuit
• JST 2M annually through IU for 5 years based on
  MOU (US half circuit)
• JST 3M annually (est.) for 5 years (JP half
  circuit)
• IU 300K annually for 5 years
• TOTAL 51.5M over 5 years for 10M NSF investment

22
TransPAC Peering (STAR TAP)

• vBNS
• Abilene
• ESNet
• NREN/NISN
• EuroLink
• MirNet

23
(No Transcript)
24
2. Allow access only to HPIIS approved
institutions

• Policy based routing restricts TransPAC traffic
  to HPIIS authorized institutions only
• AP traffic segregation done at Tokyo XP
• At STAR TAP only authorized networks can peer
  with TransPAC
• Traffic segregation is effective

25
HPIIS Networks Acceptable Use Policy

• HPIIS networks are provided for non-commercial
  advanced research and education applications.
• Limited extensions are granted to the research
  arms of commercial organizations that participate
  in collaborative projects with non-commercial
  research and education institutions.

26
(No Transcript)
27
3. Monitor the performance and use of the
TransPAC connection

• General traffic monitoring done at STAR TAP and
  Tokyo XP with MRTG and OC3Mon with graphs
  available on the TransPAC web site
• BGP session monitoring at STAR TAP
• Sysmon router exception logs available on-line
• Top talkers display (from MirNet MADAS, soon)
• Looking Glass servers available at both ends of
  the network
• Test workstations available in Chicago and Tokyo
  for special monitoring applications

28
4. Cooperate with NLANR to develop and test new
Internet protocols

• TransPAC participates in MBone, 6Bone and web
  cache projects
• TransPAC supports
• native IPv6 (application example telemicroscopy)
• native multicast (many events and meetings are
  broadcast annually)
• APAN web cache
• Production diffserv QoS in the works
• Several experiments and demonstrations in the US
  this year will help define production parameters

29
5. Provide publicly-accessible information about
TransPAC and the projects it enables

• Information avaiable through the TransPAC web
  site (www.transpac.org)
• Monthly and annual reports
• Utilization, current and historical record
• Descriptions of major projects and applications
• Network status
• AUP and application procedures
• Problem reporting and tracking
• User support contact information
• Application development and tuning information
• Engineering documents

30
(No Transcript)
31
6. In concert with NLANR provide

• Consultative user services
• DAST provides application tuning workshops for US
  TransPAC users and consulting to resolve
  performance problems.
• TransPAC engineering staff actively engaged in
  NLANR activities including participation in 1999
  and 2000 NLANR Joint Techs meetings
• NLANR operates TransPACs OC3mon at the STAR TAP
  and makes traffic data available in the DataCube
  (moat.nlanr.net/Datacube)
• Front-end tools for direct access to
  differentiated network services
• Currently not a big issue due to bandwidth
  expansions (QoS by over-provisioning)
• TransPAC is a focal point for coordination
  between I2 QBone effort and APAN QoS Working
  Group

32
Enabling Science
33
HPIIS and TransPAC Expected Outcomes

• Better infrastructure for international
  scientific collaborations
• Qualitative and quantitative changes in research
  methodologies
• New modes of communication between individuals
  and groups

34
Unexpected outcomes

• Global standardization in data handling,
  aggregation and distribution, e.g.
• HEP GriPhyN
• Genomics BioMirror and AP-BioNet
• Astronomy Sloane DSS
• Much lower barriers to the use of world class
  instruments (e.g. Osaka electron microscope)
• Greater interest in forming international
  collaborations

35
Earth ObservationData and Information Access
Link (DIAL)
DIAL is a web-based distributed system to search,
access and visualize satellite remote sensing
data for Global Change research. In collaboration
with NASDA and other institutions, NASA has DIAL
servers set up to distribute satellite remote
sensing data. NASA and NASDA also collaborate on
the Tropical Rainfall Measurement Mission (TRMM)
3D data is transferred from NASA to NASDA using
TransPAC/APAN, processed and visualized for the
web.
36
High Energy PhysicsJapan-US-Australia
Collaboration in the Silicon Vertex Detector
Project for the BELLE High Energy Physics
Experiment at KEK
The BELLE detector is the state-of-the-art
detector to investigate CP violating phenomena
with unprecedented precision at the KEK B meson
factory. The CP (CCharge conjugation, PParity)
violation is a key to explain why the universe is
dominated by the matter, not by the anti-matter.
The primary goal of the BELLE detector is to
identify the origin of the CP violation. The
BELLE collaboration consists of more than 40
institutions from Japan, Korea, China, Taiwan,
India, Russia, USA, Australia, and Europe.
37
Scientific InstrumentsTrans-Pacific
Telemicroscopy
Scientists at the Osaka University Research
Center for Ultra High Voltage Electron Microscopy
(UHVEM) and University of California San Diego
National Center for Microscopy and Imaging
Research (NCMIR) successfully use international
advanced research networks to couple the world's
largest and most powerful (3 million volt)
transmission electron microscope at UHVEM to a
remote-use computer pavilion set up at NCMIR.
http//www.npaci.edu/online/v3.10/telemicroscopy.h
tml
38
Scientific InstrumentsJapan-US Collaboration in
the Sloan Digital Sky Survey
Sloan Digital Sky Survey (SDSS) is a project to
carry out imaging and spectroscopic surveys of
half the northern sky using a dedicated,
wide-field, 2.5-m telescope. The imaging survey
with a large mosaic CCD camera will produce
digital photometric maps of the sky in five color
bands. These maps will be used to extract the
position and various photometric parameters of
about 100 million galaxies and close to the same
number of stars. The SDSS is a collaborative
project between the US and Japan involving seven
US institutions and the Japan Promotion group
(JPG). The JPG will produce merged pixel maps
from flat-fielded data.
http//www.sdss.org/
39
Distributed Computing and Data ResourcesMaximum
Likelihood Analysis Of Phylogenetic Data
DNA data has accumulated more rapidly than
compute power so researchers must often exclude
potentially infor-mative data to make statistical
analysis practical. Utilizing the computationally
intensive maximum-likelihood method of
phylogenetic inference in a globally distributed
collection of computational nodes, Indiana
University, National University of Singapore and
ACSys CRC in Australia have analyzed the DNA of
cytoplasmic coat proteins, micro-sporidia, and
cyanobacteria.
http//www.indiana.edu/rac/hpc/cp.html
40
Global Distribution of Research DataBio-Mirror
Sequence Bioinformatic data
Bio-Mirror is a world bioinformatic public
service for high-speed access to up-to-date DNA
protein biological sequence databanks. Mirror
sites have been established in Australia, China,
Japan, Korea, Singapore, Thailand, and the US.
New data are propagated to all mirror sites as
soon as they are made available, creating a
uniform, universal and reliable data base for
biological and medical research.
http//www.bio-mirror.net/
41
Advanced CollaborationSpace Physics and Aeronomy
Research Collaboratory
The Space Physics and Aeronomy Research
Collaboratory (SPARC) is an NSF-sponsored
community resource for the upper atmospheric and
space sciences operating 24 hours a day for
scientific collaboration and access to real-time
and archival data.
http//sparc-1.si.umich.edu/sparc/central
42
Conclusions

• HPIIS and TransPAC are excellent investments for
  US science that allow access to global facilities
  and expertise
• The HPIIS program has set new standards for
  infrastructure in network-enabled international
  scientific collaboration
• For many new projects this is critical
  infrastructure
• TransPAC has been able to deliver 4-5 times the
  capabilities expected for the NSFs original
  investment
• Synergies between HPIIS projects built on the
  global network crossroads, the STAR TAP, have led
  to the establishment of an integrated global NOC
  at IU