Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
IBBT Tr@ins projectTRAIN IP Network ServicesTRAIN IP Network Services
Jan Van Ooteghem & Daan Pareit
UGent – IBCN
05/11/2008
Railway Interiors Expo, Amsterdam RAI
IBBT: Demand driven ICT research
Government
Companies
new mediaeHealth
eGovernment
mobility
Government
Non-profit
enabling
technologies
Tr@ins project
� Apr 2006 – Mar 2008
� Tr@ins = TRAin IP Network Services
� Focus on
� Technical
� Broadband Internet
Research groups
� Broadband Internet
� Onboard mobility support
� Uninterrupted Internet
� Quality of Service
� Economical
� User studies
� Business model
Industrial partners
IBBT Tr@ins project
ConclusionsTechnical Business model
NAT
DHCP
AAA
Train management
NOC
Internet
Mobility Management
Local content
Mobility Management
Onboard network
� Horizontal handover
� From Access Point to Access Point
� Switch in < 50 ms for VoIP
NAT
DHCP
AAA
Train management
NOC
Internet
Mobility Management
Local content
Mobility Management
NAT
DHCP
AAA
Train management
NOC
Internet
Mobility Management
Local content
Mobility Management
Fast Roaming WLAN Switch for Rail Systems
Radius
ServerGateway
Trackside network: WLAN Fast Handover
Fast Roaming
WLAN AP
Fast Roaming
WLAN AP
Fast Roaming
WLAN AP
Fast Roaming
WLAN AP
Fast Roaming
WLAN Client
Data – 802.11a/g
Control – 802.11h
Roaming time
Average 5.64 ms
Std. dev. 0.05 ms
NAT
DHCP
AAA
Train management
NOC
Internet
Mobility Management
Mobility
Management
Local content
Mobility Management
Mobility
Management
NAT
DHCP
AAA
Train management
NOC
Internet
Mobility
Management
Local content
Mobility
Management
…
Mobility Management Central Mobility Management Train
1
…
1
M
1
Da
ta p
lan
e
N…
1
M
GatewayAggregation
unit
GatewayAggregation
unit… N MBuffer
QoS-
scheduler
Data
analysis
QoS management
…
1
M
1
…BufferQoS-
scheduler
Data
analysis
QoS management
Handover protocol
Tunnel management
Monitoring
Handover protocol
Tunnel management
Monitoring
Co
ntr
ol
pla
ne
5 QoS-classes:
• Background: FTP• Best effort: Surfing, Email• Streaming: Video stream• Interactive: VoIP call• Priority: Crew Comm.
QoS classes are mapped on outgoing links
(Load balancing)
…
Mobility Management Central Mobility Management Train
QoS management
…
1
N …
1
M
1
M
Da
ta p
lan
e
QoS management
…
1
M
1
N
…
…
1
M
GatewayAggregation
unitBuffer
QoS-
scheduler
Data
analysis
GatewayAggregation
unitBuffer
QoS-
scheduler
Data
analysis
Handover protocol
Tunnel management
Monitoring
Handover protocol
Tunnel management
Monitoring
Co
ntr
ol
pla
ne
•Decides on interfaces to be used
•Decision based on:
•SNR, packet loss, traffic measurements
•GPS, speed & management info
•IEEE 802.21 MIH events and triggers
•Load balancing, cost
…
Mobility Management Central Mobility Management Train
QoS management
…
1
N …
1
M
1
M
Da
ta p
lan
e
QoS management
…
1
M
1
N
…
…
1
M
GatewayAggregation
unitBuffer
QoS-
scheduler
Data
analysis
GatewayAggregation
unitBuffer
QoS-
scheduler
Data
analysis
•Two alternatives
•Mobile IP (MIP)
•Mobile SCTP (mSCTP)
•Multihoming
•Reliability
•Packet aggregation
•User devices don’t need to be mSCTP or MIP-aware
Tunnel management
Monitoring
Tunnel management
Monitoring
Co
ntr
ol
pla
ne Handover protocol Handover protocol
Handover protocol
20,0
25,0
30,0
35,0
40,0
Throughput(Mbps)
L2 Throughput in a WLAN
IP
SC
TP
CH
UN
K
TC
P/
UD
P Passenger
data
IP IP
SC
TP
CH
UN
K
TC
P/
UD
P Passenger
data
IP
Effective
0,0
5,0
10,0
15,0
20,0
0 500 1000 1500
(Mbps)
Passenger IP packet size (bytes)
mSCTP
MIP
IP TC
P/
UD
P Passenger
data
IP IP TC
P/
UD
P Passenger
data
IP
Handover protocol
20,0
25,0
30,0
35,0
40,0
Throughput(Mbps)
L2 Throughput in a WLAN
IP
SC
TP
CH
UN
K
TC
P/
UD
P Passenger
data
IP
CH
UN
K
TC
P/
UD
P Passenger
data
IP …
Effective
0,0
5,0
10,0
15,0
20,0
0 500 1000 1500
(Mbps)
Passenger IP packet size (bytes)
mSCTP
MIP
IP TC
P/
UD
P Passenger
data
IP IP TC
P/
UD
P Passenger
data
IP
IBBT Tr@ins project
ConclusionsTechnical Business model
Forecasting
Determining
rollout scheme
Forecasting the
number of
train passengers
Flowchart business case
Technology
assignment
Cost/benefit
model
Evaluation
Calculating
potential Internet
on train users
Determining
min. technology
requirements
Computing
technical
scenarios
Calculating
cost / benefit
model
Output business
model
Forecasting
60000
80000
100000
Nu
mb
er
of
pa
ss
en
ge
rs p
er
da
y
Relation
Oostende - Eupen
Passenger forecasting
6%
8%
10%
12%
14%
16%
18%
20%
First class adoption
Second class adoption
0
20000
40000
60000
Nu
mb
er
of
pa
ss
en
ge
rs p
er
da
y
Adoption
0%
2%
4%
6%
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Technology assignment
Technology mapping
None - UMTS – HSDPA – WiMAX
2008
2009
2010
20112011
2012
UMTS +
WiMAX
Cost parameters
Wireless
data
networks
Mobile
networks
Satellite
networks
Train equipment � Outdoor antenna
� Indoor network+ + +++
Network equipment � Trackside network
� NOC+++ + +C
ap
Ex
� NOC
General costs � Helpdesk
� Sales (billing)
� Marketing
++ ++ ++
Operations � Maintenance and repair
� Network management
� License costs
+++ + ++
Network connection � Outdoor link + +++ ++
Op
Ex
Revenue schemes
� Ticket types
� Prepaid cards per hour (e.g. hotspot service)
� Monthly subscriptions
� Ticket price
� Dependent on offered service: €2.9 – €3.9 (5 - 55 kbps)� Dependent on offered service: €2.9 – €3.9 (5 - 55 kbps)
� Service of 30 kbps: Prepaid: €3.4 / Subscription: €18
� Two revenue scenarios
� Both first and second class paying
� First class free Internet, second class paying
���� Modal switch from 1st to 2nd class!!!
NPV analysis (10 years, 15% discount rate, 1st class free Internet)
0
5
10
15
NP
V [
M€
]
UMTS
WIMAX
UMTS + WIMAX
WIMAX + UMTS
-15
-10
-5
5 10 15 20 25 30 35 40 45 50 55
NP
V [
M
Bandwidth per user [kbps]
WIMAX + UMTS
WiMAX STAT.
SAT 1 WAY
SAT 2 WAY
-5
-10
-15
30 kbps per individual user ≈ user experience of ± 1 Mbps
0
5
10
15
NP
V [
M€
]
UMTS
WIMAX
UMTS + WIMAX
WIMAX + UMTS
NPV analysis (10 years, 15% discount rate, 1st class free Internet)
-15
-10
-5
5 10 15 20 25 30 35 40 45 50 55
NP
V [
M
Bandwidth per user [kbps]
WIMAX + UMTS
WiMAX STAT.
SAT 1 WAY
SAT 2 WAY
-5
-10
-15
30 kbps per individual user ≈ user experience of ± 1 Mbps
0
5
10
15
NP
V [
M€
]
UMTS
WIMAX
UMTS + WIMAX
WIMAX + UMTS
NPV analysis (10 years, 15% discount rate, 1st class free Internet)
-15
-10
-5
5 10 15 20 25 30 35 40 45 50 55
NP
V [
M
Bandwidth per user [kbps]
WIMAX + UMTS
WiMAX STAT.
SAT 1 WAY
SAT 2 WAY
-5
-10
-15
30 kbps per individual user ≈ user experience of ± 1 Mbps
0
5
10
15
NP
V [
M€
]
UMTS
WIMAX
UMTS + WIMAX
WIMAX + UMTS
NPV analysis (10 years, 15% discount rate, 1st class free Internet)
-15
-10
-5
5 10 15 20 25 30 35 40 45 50 55
NP
V [
M
Bandwidth per user [kbps]
WIMAX + UMTS
WiMAX STAT.
SAT 1 WAY
SAT 2 WAY
-5
-10
-15
30 kbps per individual user ≈ user experience of ± 1 Mbps
0
5
10
15
NP
V [
M€
]
UMTS
WIMAX
UMTS + WIMAX
WIMAX + UMTS
NPV analysis (10 years, 15% discount rate, 1st class free Internet)
-15
-10
-5
5 10 15 20 25 30 35 40 45 50 55
NP
V [
M
Bandwidth per user [kbps]
WIMAX + UMTS
WiMAX STAT.
SAT 1 WAY
SAT 2 WAY
-5
-10
-15
30 kbps per individual user ≈ user experience of ± 1 Mbps
Sensitivity results
-94%
2%
1%
1%
-1%
-100% -50% 0% 50% 100%
Bandwidth per user
Number of SIM cards per train
Second class adoption
Modal switch
Cellular bandwidth cost
CASE 1 Mobile networks
Due to the high bandwidth
cost for UMTS
UMTS
51%
30%
-10%
4%
2%
-100% -50% 0% 50% 100%
Second class adoption
Bandwidth per user
Tariff per hour
Modal switch
Leasing percentage poles
CASE 5
Full WiMAX network
More users + more
bandwidth (less important
with dedicated networks)
= higher revenue
WiMAX
IBBT Tr@ins project
ConclusionsTechnical Business model
Summary 1
When more bandwidth is guaranteed to the customer:
� UMTS cases
� Only UMTS: Problems � No QoS
� Combination with WiMAX: QoS can be guaranteed� Combination with WiMAX: QoS can be guaranteed
� Pre-installation WiMAX: more optimized network rollout
� WiMAX case:
� Only interesting for large bandwidth
� Satellite cases:
� UMTS as gap filler: bandwidth problems
� 1Way Satellite: uplink bandwidth problems (UMTS)
� 2Way Satellite: less dependent of bandwidth
Summary 2
� Combination of technologies required
� Dense railway network: UMTS + WiMAX
� Rural railway network: Satellite solutions
� Business case
� Positive case in Belgium is possible� Positive case in Belgium is possible
� Influenced by many factors
� User forecasting, rollout scheme, offered bandwidth
� Business models
� Role and participation of the partners important
Thank you!
Contacts: