Transcript Global Terabit Research Network

Global Terabit Research Network
- Just a Dream?
Telecommunications
Portorož, 14.11.2002
Tomaž Kalin, David West
DANTE
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Issues
• Technical
– Speed on a fibre
– Switching (packet, circuit or combined)
– Multiple Lambda or Fibre (dark or lit)
• Political
– “Geopolitics”
– Financial
– Organisational
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Technical
• Speed on the fibre:
– At present 10 Gb/s/lambda (in the laboratory 80) in near
future 40 Gb/s/lambda in production
– Total capacity at this time ~ 150 Gb/s
– A single planned experiment (VLBI-E) of the order of
15 Gb/s
– GRID computing - massive data, high speed, QoS
– GÉANT will have to upgrade in the near future
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Ipv6 Land Speed Record
• Internet2 Land Speed Competition:
–
–
–
–
–
–
–
–
October 9, 2002
ARNES, RedIRIS, DANTE
Static routing, native IPv6, standard TCP/IP, single stream
Route: SI - AT - NY - UK - FR - CH - IT - ES
Length 14800 km
5154 terabit-meters/second; 384.22 Mb/s
Better than any previous single/multiple IPv4 or IPv6
Concurrently with the production traffic
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Technical
• Switching
– Present IP technology employs packet switching
– Future applications require high speed dedicated links
– Planning for a layered, mixed technology network with IP
and optical switching
• Multiple lambda, dark or lit fibre
– Dark fibre requires “man on wheels”
– Regardless of the “fibre glut” they are not available in all
places
– Stick to multiple lambda for a while
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Global Networking: some way to go
World Areas
Europe
North America
National
Networking
Established
Y
Regional Networking
Established
Well
Resourced?
Y
(GÉANT)
Y
N
Y
Y (mainly)
Asia Pacific
Y (Canada,
but several US
networks)
Y (mainly)
Central Asia
some
Partial (and
fragmented)
N
South East Europe
some
Planned (SEEREN)
N
Latin America
some
Planned
(CLARA)
some
Mediterranean
some
Planned
(EUMEDCONNECT)
N
N
Africa
N
N
N
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Horses for Courses….
• A single approach to developing global
connectivity with all regions is not
practical:
– Different stages of networking development.
– Different levels of regional organisation
– Different funding levels and structures
• Various approaches needed for greatest
effect and most cost effective
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Options
‘Come to GÉANT’
– Euro-centric approach, offers transit within and across the EDA, would lose/slow
most inter-regional connectivity if sole approach
• ‘Co-operative development’
– The GTREN approach, fairest but slow, cumbersome and potentially complex to
administer. Embryonic networks not ready for this
• ‘Meet in the US’
– The historical approach, gave connection to all established networks, but unequal
costs, doesn’t help develop embryonic networks
• Inter-regional consortium backbone’
– GTREN owned by consortium of regional networks, complex and politically
difficult to set up, shares costs
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
‘Come to GÉANT’ Option
• Most applicable for Central Asia, Asia Pacific and
Mediterranean, also relevant to Latin America and other
regional networks
• Gives Connectivity across GÉANT EDA to all European
NRENs and other regional networks
• NREN or regional entity provide connection to EDA
• Termination on GÉANT router
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Co-operative Development Option
• Most applicable to North America:
– Transatlantic connectivity provided by GÉANT:
7.5Gb/s
– Connectivity provided by North America:10 Gb/s ????
• Problems by North American funding issues,
driven by IEEAF donations
• Co-operative approach is complex to manage but
worth continuing
• Proposed bilateral collaboration with Canarie
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
GÉANT Inter-Regional Backbone Option
• Most applicable to:
– Asia Pacific 2003 (proposed in TEIN)
– Latin America 2004 (already planned through ALIS)
– Mediterranean 2005?
• Extends GÉANT backbone into these regions
building on transatlantic experience
• Invite in-region NRENs to connect at own
expense to the nearest regional GÉANT PoPs
• As regional networking develops, seek cost
sharing
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Financial
• Costs still significant:
–
–
–
–
Telcos need income - not reducing prices, willing to upgrade
Cost of routers becoming a problem
Cost of local tails !
Problem of cost sharing with the US networks a serious
issue
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West
Conclusion
• GTRN - a dream?
– Yes, a dream in short term sense, if speaking of networks
with trunks running at terabit speeds
– Not a dream if speaking of the global connectivity
– Not a dream if speaking of backbone networks with
aggregate speeds of multiple terabits per second
• Timescale - 3 to 4 years
Global Terabit Research Network - Just a Dream? - Tomaz Kalin, David West