Transcript Internet2 Organizational Update Renee Frost Ted Hanss

Optical Futures: high-performance networking
for research and education and optical
technologies
Heather Boyles
Director, International Relations
Internet2
[email protected]
21 March 2017
18 February 2003
Hong Kong
Internet2 Mission and Goals
Develop and deploy advanced network
applications and technologies,
accelerating the creation of tomorrow’s
Internet.
• Enable new generation of applications
• Create leading edge R&E network capability
• Transfer technology and experience to the global
production Internet
2
Internet2 Areas of Work
Advanced Applications
Middleware
Network Engineering
• End to End Performance
Advanced Network Infrastructure
Partnerships and Outreach
3
Internet2 Network Architecture
GigaPoP
One
GigaPoP
Two
Internet2
Backbone Networks
GigaPoP
Three
GigaPoP
(n)
4
Internet2 Network Architecture
Internet2 Backbone
Network(s)
University A
GigaPoP
One
Regional Network
University B
University C
Commercial
Internet
Connections
5
Abilene Network
Core Map, January 2003
6
Abilene Network
Logical Map
7
What is optical networking?
Utilizing optical fiber to carry light as the
lowest level data transport medium
Good source for basics of optical
networking: http://www.sura.org/opcook
8
How is optical networking being discussed in
research networking community?
Currently, two main threads of activity:
1) Obtaining and controlling fiber assets in
order to build networks (at campus,
metropolitan, regional, national levels)
2) Utilizing that control of optically-based
transport layer to provision new types of
services – “lambdas**” or “lightpaths”
….and sometimes simulating these without underlying control
**”a pipe where you can inspect packets as they enter and when they
exit, but principally not when in transit” (C. deLaat et al.)
9
How have we provisioned
networks in the past?
To date, primarily by buying services
from telecommunications companies
• ATM service, SDH/SONET service,
GigabitEthernet Services, wavelength services
Example: Abilene backbone network
• Qwest provides 10Gbps wavelength between
core nodes
• Abilene owns, controls routers
10
Providing network infrastructure
by acquiring fiber assets
Unique window in time for fiber assets
• Cause: fiber glut, bankruptcies and telcos in distress
• Within a year, opportunity on national scale closes?
• Hedge against a regression to ‘bad old days’ of monopolies
Technically, getting fiber means controlling the
network down to layer 1 (0?)
• Would allow deployment of different wavelengths for
differentiated networks (high perf advanced services,
network research, more general EDU access)
• Path to doing optical switching when it makes sense
11
Unique optical requirements in
Higher Education Community
10-Gbps: 10 Gigabit Ethernet preferred over
OC-192c SONET
HPC could need 40-Gbps λ’s prior to the
carriers
Integrated view of network management
• Transport & IP engineering/operational approaches are not
intrinsically different
• SNMP preferable for network polling
HEC can provide experimental environment
for development of ‘rational’, customer
focused optical switching
• Switching tightly integrated with optical transport
• Capacity for IP backbone expansion and p2p λ’s
12
Metro/regional implementations lead
(for technical and economic reasons)
Distance
scale (km)
Examples
Equipment
UW(SEA),
Metro
< 60
State/
Regional
< 500
Dark fiber & end
terminals
USC/ISI(LA)
I-WIRE (IL),
Add OO
CENIC ONI,
amplifiers
I-LIGHT (IN)
PLR,
Extended
Regional/
National
> 500
TeraGrid
Add OEO
Abilene
regenerators
National Light Rail
& O&M $’s
Variety of campus, metropolitan,
regional optical networks in US
CALREN2 – California
IWIRE – Illinois
ILIGHT – Indiana
SURA Optical Cookbook examples
• Harvard Joint Trench Project
• Southern Crossroads (Atlanta-area dark fiber
buildout)
14
ORION Network – Overview
a Canadian (Ontario) example
Note.-This overview includes ORION PoP’s, associated RANs, and other ORANO members. Other members could be 50 – 60 R&E sites to connect to the nodes
sureNet
LaurentianU
OMAN
CambrianC
CBoreal
LakeheadU
CarletonU
UOttawa
NorthernC
AlgonquinC
LaCiteC
ConfederationC
NipissingU
LaurentianU
SaultC
CanadoreC
LambtonC
GeorgianC
ConestogaC
UWaterloo
WLU
UGuelph
SSFlemingC
CentennialC
CGrandLacs
GBrownC
HumberC
SenecaC
TrentU
LoyalistC
Queen'sU
RMC
SLawrenceC
DurhamC
SheridanC
LARG*net
UWO
FanshaweC
McMasterU
OCAD
RyersonPU
UToronto
YorkU
MohawkC
BrockU
City / Town with
University only
City / Town with
College only
WEDnet
City with CA*net 3
GigaPoP
UWindsor
St.ClairC
LEGEND
City / Town with
College and University
College [25]
University[19]
Potential
ORION
backbone
NiagaraC
Source: Randy Neals, ORANO
15
Regional optical fanout
In the next generation architecture,
regional & state based optical
networking projects are critical
Three-level hierarchy remains vital
• National backbone, GigaPoPs (ARNs),
Campuses
Close collaboration with the Quilt
GigaPoPs
• Regional Optical Networking project
16
Toward a National Optical
Networking Facility
Research and education community
investment in national-scale fiber assets
• Discussions among a number of partners in US
ongoing
– “National Light Rail” – being led by members of
Internet2 community – CENIC, the Pacific NorthWest
Gigapop and other partners
– SURA – USAWaves project
17
CA*net 4 Architecture
source: Bill St. Arnaud
Edmonton
Saskatoon
Winnipeg
Vancouver
Calgary
Kamloops
Regina
Halifax
Thunder Bay
St. John's
Victoria
Quebec City
Sudbury
Seattle
Charlottetown
Montreal
Ottawa
Minneapolis
CA*net 4 Node
Fredericton
Toronto
Buffalo
London
Possible Future Breakout
Possible Future link or Option
Halifax
Kingston
Hamilton
Boston
Albany
Windsor
Chicago
New York
CA*net 4 OC192
18
Light Path Scenarios
Workstation to Workstation Wavelength
CWDM
GigaPOP to GigaPOP Wavelength
Regina
BCnet
Vancouver
Campus
OBGP
switch
Winnipeg
St. John’s
RISQ
Halifax
Calgary
Montreal
Seattle
Toronto
Lambda* or Lightpath Networks
Current CA*net “customer-empowered
networking” - prototypes a day when multiple
wavelengths available to a site/desktop
Why?
• Router limitations - cost
• A few very bandwidth needy applications (e.g. between
radio astronomy sites) for which dedicated circuits make
more (technical, economic) sense
• User control – configurability
An area needing more investigation, cost
modeling, prototyping, etc.
20
Summary
There are many networks out there
pursuing the dark fiber opportunity
• Scale is important
• Regulatory factors
• Take advantage of others’ experiences!
The “lightpath” model will be exciting
area of investigation
• What are the right models? “Empowered
customer” pays traditional circuit-switched
network-type fees?
21
www.internet2.edu
22