Global Highperformance Networking and TransPAC2

1
Global High-performance Networking and TransPAC2
James Williams Indiana University TransPAC2
Principal Investigator williams_at_indiana.edu
2
Topics To Be Discussed

• TransPAC2 overview
• TransPAC2 architecture and engineering goals
• TransPAC2 service goals
• 2005 plan
• Coordination opportunities
• Asian politics

3
The TransPAC2 Project

• Follow-on to the TransPAC project
• Funded by the US National Science Foundation
  (SCI-0441096)
• US partners are Indiana University and Internet2
• Asian partner is APAN
• About 1M/year in funding for 5 years
• Encourages scientific collaboration between
  researchers in Asia and the US

4
TransPAC2 Architectural Goals

• Continue high-performance connectivity across the
  Pacific Ocean OC-192 connection between US and
  Tokyo
• Enhance connectivity by assisting in the
  development of an intra-Asia backbone Tokyo-Hong
  Kong-Singapore or beyond

5
TransPAC2 Architecture
6
TransPAC2 Engineering Topology

• TransPAC2 OC-192 connects APAN Tokyo XP with
  TransPAC2 router in Los Angeles
• LA TransPAC2 router maintains BGP peering with
  North American RE Networks
• Connectivity in US via Pacific Wave Ethernet
  Exchange Point
• Pacific Wave nodes in LA and Seattle
  interconnected via 10GigE lambda on National
  LambdaRail facility
• Tokyo-Hong Kong connectivity via OC-48 circuit
• Cisco 15454 SONET switches multiplex two GigE
  circuits to provide connectivity to both Hong
  Kong Exchange Points
• Hong Kong-Singapore link details to be worked out

7
TransPAC2 Engineering Topology Diagram
8
North America-Asia Circuit Priorities

• Three major production circuits between North
  America and Asia (not including AU)
• TransPAC2 (LA and Seattle)
• JGN2 (Chicago)
• NII/SINET (New York)
• All three circuits will back each other up
• Primary circuit TransPAC2
• Secondary circuit NII/SINET
• Tertiary circuit JGN2
• BGP routes will be localprefd according to the
  above scheme on all networks

9
TransPAC2 Service Goals

• Provide production networking to facilitate
  research cooperation between the US and Asia
• Cooperation with Internet2 HOPI project
• US-Asian cooperation on lightpath technologies
  will be crucial in future development
• Initial assets will allow for MPLS and
  GMPLS-like experimentation, with future
  equipment possibly supporting layer 2 and
  layer 1 services
• Security coordination
• Measurement coordination
• AAI development

10
HOPI and TransPAC2

• The Internet2 HOPI project will attempt to
  provide both routed and circuit-based
  functionality
• Over the next year, the HOPI Design Team will
  test and evaluate vendor equipment in an effort
  to define the capabilities and implications of a
  hybrid network
• The initial focus will be on lightpath creation
• In 2005, a HOPI testbed will be created to answer
  some core questions
• What defines a deterministic path through a
  layer3 network?
• latency and jitter guarantees? bandwidth
  protection?
• What does the network need to do to provide a
  lightpath
• MPLS and QoS and STS/STM and ???
• What flavor of MPLS? (L2VPN or L3VPN or stitched
  LSPs?)
• How dynamic do lightpaths need to be?
• How will the lightpaths be created?
• GMPLS? UCLP? Human?
• How will lightpaths cross administrative
  boundaries?
• Indiana University has been involved with the
  HOPI project from its inception
• TransPAC2 will work with Internet2 to develop
  mutual goals and similar resources
• Lightpath support will be important to future
  network growth, though a shared and production
  packet-based service offering will likely always
  service the majority of users
• Lightpath-based experimentation will only be
  undertaken if it will pose no significant risk to
  the production nature of the TransPAC2 service
  offering.

11
Security

• An important issue that needs IRNC-wide attention
• Need for central data collection, contact list
  and incident coordination
• Need for international incident response
• URL http//www.ren-isac.net/

12
TransPAC2 Measurement

• The TransPAC2 measurement effort will provide
  valuable data and tools to network engineers and
  researchers in order to debug network problems
  and improve application performance
• TransPAC2 will deploy an Asian Observatory
  framework in Asia, modeled after the US-based
  Abilene Observatory.
• This framework will help determine performance
  characteristics of the complete network path by
  aggregating information about the segments that
  make up the network path

13
TransPAC2 Measurement (cont.)

• The Asian Observatory system consists of PCs at
  key locations along the network backbone
  collecting and analysing measurement data
• This work makes use of existing tools such as the
  bwctl and owamp codes developed by the Internet2
  piPEs project.
• Other tools from the Abilene NOC, APAN NOC JP,
  and NLANR will be deployed on this hardware as
  well
• Data collection and analysis will conform to the
  guidelines developed by Matt Zekauskas and Matt
  Mathis for the IRNC project , the GGF, and other
  policy makers in the measurement community

14
2005 Measurement Goals

• Measurement machines will be deployed in the
  TransPAC2 US co-location space to collect data
  (some resources already located in Tokyo)
• Full code implementation of existing measurement
  and analysis tools
• Schedule persistent tests between APAN/TransPAC2
  and Abilene Observatory nodes
• Make measurement data available to the network
  and research communities
• Foster collaboration between the APAN
  measurement community and other global network
  measurement projects such as the IRNC initiative

15
Authentication and Authorization Infrastructure

• TransPAC goal to promote development of national
  AAI in APAN countries
• Why?
• In support of
• Access to lightpath/HOPI-like services
• Access to performance measurement/monitoring
  tools and services
• Authenticated sharing of network security
  information
• Inter-institutional collaboration broadly Grid,
  authenticated video-conferencing, etc.

16
AAI(2)

• How?
• Leverage NSF NMI program
• Leverage Internet2 and other countries emerging
  national AAI efforts
• Cotswolds group Cookbook
• PRAGMA collaboration

17
Supporting Production Science

• High-performance bandwidth across the Pacific
  OC-192 combined with development of an
  inter-Asia backbone provide necessary
  infrastructure for e-science
• Typical (everyday, non-heroic) applications
  include e-VLBI data transfers, HEP data transfers
  between Japan and the US, grid computing research
  and standard everyday science activities

18
Supporting Production Science (cont.)

• Physics www.riken.go.jp and www.kek.jp
• Nanotechnology www.nims.go.jp
• Grid computing www.pragma-grid.net
• Earth science www.ees.hokudai.ac.jp
• PRAGMA www.pragma-grid.org

19
TransPAC2 Current Status

• TransPAC2 has received NSF funding (SCI-0441096)
• James Williams, Indiana UniversityPrincipal
  Investigator
• Douglas Van Houweling, Internet2Co-PI
• APAN network owners committed to the TransPAC2
  effort
• Purchase of OC-192c services complete
• OC-48 between Tokyo and Hong Kong in place
• Purchase of US end equipment to be located in Los
  Angeles is complete
• We expect the OC-192 to be active 5/1/2005

20
2005 Plans

• Operationalize OC-192 May 1
• Establish link to Singapore Oct 1
• Web site May 1
• Coordinating Committee May 1
• Measurement mentioned previously
• Security being discussed
• HOPI ongoing
• AII ongoing

21
Coordination Opportunities

• Sharing of bandwidth across the Pacific (10Gbp is
  plenty for all)
• Linking AU and Hawaii into Asia in a cost
  effective manner
• Clarifying the Hong Kong route exchange issue and
  general HK connectivity
• Exchange Point pricing

22
Interesting Networking Observations

• The bandwidth problem between the US and Asia
  has been solved! gt60Gbps available as of 4/1/05
• The access problem within Asia continues
• Tien2 Project is one effort to address this
  problem
• The north-south problem within Asia continues

Now we have to address the network service
infrastructure and scientific collaboration
problem.
23
Interesting Asian Political Issues

• Politics and economics are now allowing more
  networking cooperation among Asian countries
• Regulated telecommunications environments
• Desire to connect directly to the US
• Uncomfortable nature of sharing
• Within country disputes

24
James Williams TransPAC2 Principal
Investigator williams_at_indiana.edu