Transcript 20010726-Engineering-Almes

Internet2 Engineering Issues
IBM T J Watson :: Hawthorne
Guy Almes <[email protected]>
25 July 2001
Internet2 Engineering
Objectives
A
Provide our universities with superlative
networking:
• Performance
• Functionality
• Understanding
A
Make superlative networking strategic for
university research and education
Engineering:
Advanced Functionality
A
Multicast
A
IPv6
A
QoS
A
Measurements
A
Support for End-to-End Performance
Internet2 Multicast
A
Multicast Working Group
• Kevin Almeroth, Univ California Santa Barbara, chair
A
Encouraging more pervasive high-quality
deployment of native IP multicast
throughout the Internet2 infrastructure
A
Fighting fires
A
Keeping an eye on SSM
A
Clarifying the application story
Internet2 Multicast
Architecture
A
PIM-SparseMode
• multicast routing within an Autonomous System
• quite scalable
• notion of rendezvous points
A
MBGP
• between Autonomous Systems
A
MSDP
• Source Discovery
Longer-term WG Issues
A
A
Scalability (what happens if it does catch
on?)
Exploring the role of Source-Specific
Multicast
Could SSM be Enough?
A
'Classic' Multicast
• Group <g> has global significance
• A user creates, joins, sends to g
• Others can join, then send to and/or listen to g
• MBGP, PIM-SM, MSDP triad
A
Source Specific Multicast
• Group <g> has local significance
• A user 's' creates, sends to <s,g>
• Others can subscribe to, then list to <s,g>
• No need for MSDP (or allocation of <g> values)
Implications of SSM
A
Simplify Multicast Routing / Addressing
• No need for global class-D address allocation
• No need for source discovery
A
Complicates 'few-to-few' applications
• Define all the members of the application-level group
• Both a burden and an opportunity
A
Allows better Security, Scalability
A
Requires new version of IGMP
Multicast Summary
A
Full functionality supported now
A
Deployment steadily increasing
A
Some international peering, e.g., CA*net3
A
Performance excellent
A
Scalability?
A
Applications?
Internet2 IPv6
A
IPv6 Working Group
• Dale Finkelson, Univ Nebraska, chair
A
A
A
Build the Internet2 IPv6 infrastructure
Educate campus network engineers to
support IPv6
Explore the Motivation for IPv6 within the
Internet2 community
IPv6 Infrastructure
A
vBNS and Abilene both support IPv6
A
Abilene IPv6 with IPv6/IPv4
• Four 'backbone' nodes: Cisco 7200
"
Atlanta, Pittsburgh, Denver, and Indianapolis
• Managed by the Abilene NOC
A
IPv6 WG: address allocation and
engineering coordination
Education / Training
Goals
A
IPv6 hands-on workshop
• Lincoln, Nebraska; 17 May 2001
• starting from scratch, build an IPv6 network,
including routers, hosts, DNS tools and various
transition tools, ending up with a functional IPv6
network fully interconnected to the global Internet.
A
Materials from this workshop will be
available to enable gigaPoPs and others to
use in their own workshops.
Explore IPv6 Motivation
A
Why should our users, campus decisionmakers, and community generally care
about IPv6?
• IPv6 preserves the classic end-to-end transparency of
the Internet architecture
• improved support for mobility
• key for IPsec
• key for the scalability of the Internet
A
The answers must be pragmatic.
Engineering:
End-to-End Performance
The Current Situation
A
Our universities have access to an
infrastructure of considerable capacity
• examples of 240 Mb/s flows
A
End-to-end performance varies widely
• but 40 Mb/s flows not always predictable
• users don't know what their expectations should be
A
Note the mismatch
Threats to
End to End Performance
A
A
BW = C x packet-size / ( delay x sqrt(packet-loss ))
(Mathis, Semke, Mahdavi, and Ott, CCR, July 1997)
Context:
•
Network capacity
•
Geographical distance
•
Aggressive application
Threats to
End to End Performance
A
Fiber problems
• dirty fiber
• dim lighting
• 'not quite right' connectors
Threats to
End to End Performance
A Fiber problems
A
Switches
• horsepower
• full vs half-duplex
• head-of-line blocking
Threats to
End to End Performance
A Fiber problems
A Switches
A
Inadvertently stingy provisioning
• mostly communication
• happens also in international settings
Threats to
End to End Performance
A Fiber problems
A Switches
A Inadvertently stingy provisioning
A
Wrong Routing
• asymmetric
• best use of Internet2
• distance
Threats to
End to End Performance
A Fiber problems
A Switches
A Inadvertently stingy provisioning
A Wrong Routing
A
Host issues
• NIC
• OS / TCP stack
• CPU
Perverse Result
A
A
'Users' think the network is congested or
that the Internet2 infrastructure cannot
help them
'Planners' think the network is
underutilized, no further investment
needed, or that users don't need high
performance networks
Promising Approaches
A
Work with key motivated users
A
'Shining a flashlight' on the problem
A
Measurements
A
Divide-and-Conquer
A
Understanding Application Behavior
A
Getting it right the first time
Internet2 End-to-End
Performance Initiative
A
Very recently hired / deployed staff
• Cheryl Munn-Fremon, initiative director
• Russ Hobby, chief technical architect
• George Brett, chief information architect
A
$1.5M budgeted by Internet2
Internet2 End-to-End
Performance Initiative
A
Distributed measurement infrastructure
• Enable rapid effective understanding of why an
instance of end-to-end performance is limited
• Make the work of PERF participants rewarding
• Enable initiation of tests by PERF participants
A Teams of performance analysis specialists (PERF)
A Dissemination of best practices
Internet2 End-to-End
Performance Initiative
A Distributed measurement infrastructure
A
Teams of performance analysis specialists
(PERF)
• members at campuses, gigaPoPs, backbones
• socially and technically coordinated
• committed to effecting radical change
A Dissemination of best practices
Internet2 End-to-End
Performance Initiative
A Distributed measurement infrastructure
A Teams of performance analysis specialists (PERF)
A
Dissemination of best practices
• Identify key techniques, tools, and 'best practices'
• Make them common
• Work toward widespread / routine excellent user
experiences
• Improve the reputation / status of network engineers
Anticipated Partners
A
NLANR: DAST, MOAT, and NCNE
A
Web100 Project
A
Abilene partners
A
Leading campuses and gigaPoPs
A
Internet2 corporate members
Access to Key Resources
A
Optical telescopes in Hawaii
A
CRAFT Project
A
PACI Supercomputer Facilities
A
CERN