Transcript Making the Wireless Internet Possible… and Profitable

IP Core Networks for
cdma2000 Radio Access
David Boettger
CDMA Solutions Marketing
Hong Kong
3G Conventional Wisdom
• 3G discussions often revolve around
– Radio access (especially speed)
– Wireless Internet services
• Limited discussion about 3G costs
…But Immediate Operator Concerns
Surround Costs and Margins
• Falling per-minute prices due to competition
• Rising operational costs with subscriber growth
• Requirement for low-cost network evolution in
support of
– Increasing amounts of voice traffic
– New data services
cdma2000 Cost Proposition
• Well cost-reduced terminals (4+ years)
• Mature & cost-reduced radio access network
– 1X/3X scalability
• IP-enabled core network (CN)
Why IP for cdma2000 CN? (1)
• IP component hardware costs
– Unit costs much lower than existing CN nodes
– Shipment volumes dramatically higher than
existing CN nodes (“economies of scale”)
– IP components often more scaleable
* Less “lumpiness” in CN CAPEX
• Time to market
– IP component feature inheritance from other IP
product development efforts
– Open interfaces based on Internet standards
broaden supplier base
Why IP for cdma2000 CN? (2)
• Future resistance
– Computer-like performance curve of IP nodes
– IP is the protocol of choice for all new interfaces,
services, and applications
– Tendency toward IP for all networks
* Wireline, enterprise, etc.
Building the All-IP cdma2000 CN
The Managed IP Core
Managed Core Packet Network
• Policy Management
• Bandwidth Control
• Route Optimization
• Privacy
• Quality of Service
Routers, IP
Switches, Policy
Servers, Firewalls,
etc.
Building the All-IP cdma2000 CN
Delivery of Voice and Wireless Internet
PSTN
Trunks
IP
Wireless Operator
Managed Packet
Network
cdma2000
RAN
Packet Packet Data Serving Node
• Gateway between wireless network and
ISPs & corporate intranets
IP
PDSN
ISP/Corp
Building the All-IP cdma2000 CN
Tandem Traffic Migrated to Managed Packet Core
PSTN
Trunks
IP
IP
IP
IP
Wireless Operator
Managed Packet
Network
cdma2000
RAN
Packet Voice Gateway (PVG)
• Converts voice trunks to packets
• Reduces network bandwidth requirements
• Target: Lower tandem trunking costs by 2/3
IP
PDSN
ISP/Corp
PVG Tandeming Costs and Savings
$1,800,000
$1,600,000
Yearly Savings and Costs of
Packet Backbone*
$1,400,000
$1,200,000
$1,000,000
$800,000
Transport Savings with PVG
Additional Network Costs with PVG
$600,000
$400,000
$200,000
$2000
2001
2002
2003
2004
$250,000
Cost Breakdown of Implementing
Packet Voice Gateway*
$200,000
$150,000
PVG Equipment Maintenance
PVG Installation & Commissioning
$100,000
$50,000
$2000
2001
2002
2003
*Based on a network with 500,000 subscribers growing to 1 million in 2004.
2004
Building the All-IP cdma2000 CN
Full IP Evolution
Elimination of circuit paradigm
• Off-the-shelf IP/ATM
switches, routers, and
gateways
• Call and mobility services on
industry standard computers
PSTN
Call Server
Wireless Operator
Managed Packet
Network
cdma2000
RAN
IP
SG
SS7 Network
PDSN/
GGSN
ISP/Corp
All-IP CN Case Study
Nortel Networks Commercial cdmaOne Systems
• Compare All-IP CN vs. “legacy” CN costs in real, commercial
cdmaOne deployments
– Cluster of markets (300k growing to 1.6M subs)
– Smaller single market (50k growing to 285k subs)
• Answer strategic questions
– Is IP more cost effective than current CN?
– Sensitivities to
* Voice/data traffic mix?
* Market size?
– Scalability of IP CN?
– Optimal transition strategy?
Results – Initial Network Buildout Costs
300k Subscribers in 2000
25.0
Initial Build Capital Costs Year 2000:
1 Usage
ScenarioScenario
1 – Low PSD
21.6M
20.0
Real Estate
Data Netw orking
4.7
BSC/WAG function
Cost (Millions)
0.4
Sw itching Function
15.0
6.7
-23%
10.0
6.7M
5.0
0.5
0.2
9.8
-92%
5.2
0.8
PMOH/W
Current
All IPNGWN
CN H/W
• 69% All-IP CN savings over current system (63% ex. real estate)
• Given that data traffic is relatively insignificant at launch, All-IP CN clearly attractive
for voice traffic
• Reason: Dramatically lower-cost hardware components for All-IP CN
• Switching/HLR functions 92% cheaper in All-IP CN due to concentration & hardware
costs
Results – Cost of Ownership
1.6M Subscribers in 2008
120.0
PV of All Network Costs 2000-2008
Scenario Scenario
1 – Low PSD1 Usage
120.0
PV of All Network Costs 2000-2008
2 Usage
Scenario Scenario
2 – High PSD
105.1M
80.0
91.8M
BSC/WAG function
30.7
40.0
Sw itching Function
2.1
60.0
Transport
100.0
Data Netw orking
Data Netw orking
51.6M
32.5
20.7
0.7
20.0
26.5
0.0
PMOH/W
Current
PV of Costs (Millions)
PV of Costs (Millions)
100.0
Transport
37.6
Sw itching Function
1.5
62.5M
60.0
38.3
29.7
40.0
1.8
26.4
20.0
3.8
0.0
All IPNGWN
CN H/W
BSC/WAG function
80.0
27.7
26.3
4.7
PMO H/W
Current
All IPNGWN
CN H/W
• All-IP CN achieves 44% savings ($40M in PV) over current network in Scenario 1,
41% ($43M in PV) in Scenario 2
• Savings dominated by voice traffic impact
• Significant data traffic in Scenario 2 did not change fundamental rates of savings
• Reasons for savings: lower hardware costs for All-IP CN and lower transport costs due
to packetized voice traffic
All-IP CN Cost Savings Analysis
• All-IP CN provides significant savings in initial
network build and total cost of ownership
• Savings realized through reduced transport OPEX
and hardware CAPEX
• Voice traffic primary means of savings
– cdmaOne/cdma2000 variable-rate codec
• Savings scale with market size
Summary and Conclusions
1. IP-based cdma2000 CN allows dramatic savings
in cost-of-ownership
2. cdma2000 provides the most cost-reduced and
mature 3G RAN
3. When combined with the most cost-reduced
handsets…
cdma2000 is the lowest-cost
path to 3G services