multi-service backbone design drivers behind next generation networks vijay gill jim boyle
TRANSCRIPT
Multi-Service Backbone Design
Drivers behind Next Generation Networks
Vijay Gill <[email protected]>Jim Boyle <[email protected]>
What
mul·ti·serv·ice adj.
Offering or involving a variety of services
IP, Voice, Private Line, VPNs, ATM, FR
Why
$700 Billion – Voice, Fax, Modem Market
Telco Companies Day Job Voice based communications is still
~ 90% of Global Telco revenue Voice bit is ~ 14 x more expensive
than data bit
Ways to Deliver Multiple Services
Multiple Backbones One for each Service (SONET, ATM/FR,
IP) Common Backbone
Layer multiple services on top of a common transport fabric
Multiple Backbones
Application Aware Dedicated Infrastructure used to implement
each application – PSTN, FR, ATM, Private Line Discourages Diversity
Needs Large Market Demand before it is cost effective to go out and build the support infrastructure
ATM/SONET infrastructure More boxes, complex management issues Hard to upgrade bandwidth in sync for all backbones
Common Backbone
Application Unaware Characterized by the new breed of Telcos
Rapid innovation Clean separation between transport, service, and
application Allows new applications to be constructed without
modification to the transport fabric.
Less Complex (overall)
Why A Common Backbone?
Spend once, use many Easier capacity planning and implementation
Elastic Demand 1% price drops result in 2-3% rise in demand –
Matt Bross, WCG Increase of N on edge necessitates 3-4 N core
growth Flexibility in upgrading bandwidth allows you to
drop pricing faster than rivals
Source: KPCB
Historical and forecast market price and unit cost of Transatlantic STM-1 circuit (on 25 year IRU lease)
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Pri
ce p
er S
TM
-1 (
$m)
PRICE
COST
Bandwidth Blender - Set on Frappe
Some Facts
There is no absolute way to measure any statistic regarding the growth of the Internet
The Internet is getting big, and it's happening fast
Source: Robert Orenstein
Solution
Leverage packet based technology Multi service transport Fabric Optimize for the biggest consumer - IP
Provide a loosely coupled access point for service specific networks (e.g. IP good, per-call signaling bad)
Solution
Internet (IP) Internet (IP)
VPN VPN
Voice/Video
CES
Voice/Video
CES
Multi Service IP Transport Fabric
Requirements
Isolating inter-service routing impacts
Address space protection/isolation Fast Convergence (Service
Restoration) Providing COS to services
Requirements
Support multiple services Voice, VPN, Internet, Private Line
Improving service availability with stable approaches where possible
LSPs re-instantiated as p2p links in IGP e.g. ATL to DEN LSP looks like p2p link
with metric XYZ Run multiple instances of IGPs (TE
and IP)
Stabilize The Edge
Stabilize The Core
Global instability propagated via BGP Fate sharing with the global Internet
All decisions are made at the edge where the traffic comes in
Rethink functionality of BGP in the core
COS
Mark service bits upon ingress WRR on trunks configure max time-in queue Avoid congestion But when congested, monitor that
traffic delivered in line with objectives Crude (compared to what?) but
effective.
Implementation Approaches
Pure IP Layer 2 tunneling (aka CCC, AToM) RFC2547 (base and bis) Merged IGP Multi process IGP
IP + Optical Virtual Fiber Mesh Protection GMPLS (UNI, NNI)
IP Only
Fiber
DWDM / 3R
IP / Routers
• Removal of an entire layer of active optronics
• Directly running on DWDM• Technology for Private
Lines and Circuit Emulation isn’t here yet
• Fate sharing with Global Internet
LSP Distribution
LDP alongside RSVP Routers on edge of RSVP domain do fan-out Multiple Levels of Label Stacking Backup LSPs
Primary and Backup in RSVP Core Speed convergence Removes hold down issues (signaling too fast in a
bouncing network) Protect path should be separate from working
There are other ways, including RSVP E2E
IP + Optical
Fiber
DWDM / 3R
IP / Routers
Optical Switching
• Virtual Fiber • Embed Arbitrary fiber
topology onto physical fiber.• Mesh restoration. • Private Line• Increased Velocity of service
provisioning• Higher cost, added complexity
IP + Optical Network
Big Flow
Big Flow
Out of port capacity, switching speeds on routers? Bypass intermediate hops
Dual Network Layers
Optical Core (DWDM Fronted by OXC) Fast Lightpath provisioning Attach Metro collectors in Mega PoPs via multiple OC-
48/192 uplinks Metro/Sub-rate Collectors
Multiservice Platforms, Edge Optical Switches Groom into lightpaths or dense fiber. Demux in the PoP (light or fiber)
Eat Own Dog Food Utilize customer private line provisioning internally to
run IP network.