IBBT Tr@ins project TRAIN IP Network Services Jan Van Ooteghem & Daan Pareit UGent IBCN 05//2008 Railway Interiors Expo, Amsterdam RAI
IBBT: Demand driven ICT research ehealth new media Companies Government Non-profit enabling technologies egovernment mobility
Tr@ins project Apr 2006 Mar 2008 Tr@ins = TRAin IP Network Services Focus on Technical Broadband Internet Onboard mobility support Uninterrupted Internet Quality of Service Economical User studies Business model Research groups Industrial partners
IBBT Tr@ins project Technical Business model Conclusions
AAA DHCP NOC Train management NAT Mobility Management Internet Mobility Management Local content
Onboard network Horizontal handover From Access Point to Access Point Switch in < 50 ms for VoIP
AAA DHCP NOC Train management NAT Mobility Management Internet Mobility Management Local content
AAA DHCP NOC Train management NAT Mobility Management Internet Mobility Management Local content
Trackside network: WLAN Fast Handover Radius Server Gateway Fast Roaming WLAN Switch for Rail Systems Roaming time Average 5.64 ms Control 802.h Fast Roaming WLAN AP Std. dev. 0.05 ms Fast Roaming WLAN AP Data 802.a/g Fast Roaming WLAN AP Fast Roaming WLAN AP Fast Roaming WLAN Client
NOC Train management AAA NAT Mobility DHCP Management Internet Mobility Management Local content
AAA DHCP NOC Train management NAT Mobility Management Internet Mobility Management Local content
Mobility Management Central Mobility Management Train QoS management Data analysis N QoSscheduler M Buffer M Aggregation unit Gateway Data plane Gateway Aggregation unit M QoS management Buffer M QoSscheduler N Data analysis Control plane 5 QoS-classes: Tunnel management Tunnel management Background: QoS classes FTP are mapped on outgoing links (Load balancing) Streaming: Video stream Interactive: VoIP call Monitoring Priority: Crew Comm. Handover protocol Best effort: Surfing, Email Handover protocol PDF Monitoring
Mobility Management Central Mobility Management Train QoS management Data analysis N QoSscheduler M Buffer M Aggregation unit Gateway Data plane Gateway Aggregation unit M QoS management Buffer M QoSscheduler N Data analysis Tunnel management Tunnel management Decides on interfaces Handover to protocol be used Decision based on: SNR, packet loss, traffic measurements GPS, speed & management info IEEE 802.2 MIH events and triggers Load balancing, Monitoring cost Control plane Handover protocol PDF Monitoring
QoSscheduler Two Mobility alternatives Management Central Mobile IP (MIP) Mobile QoS management SCTP (msctp) Multihoming Data QoSscheduler M Reliability analysis N Packet aggregation Data plane Gateway Buffer Aggregation unit User devices don t need to be msctp or MIP-aware M M Mobility Management Train Aggregation unit QoS management Buffer M Gateway N Data analysis Control plane Tunnel management Handover protocol Monitoring Tunnel management Handover protocol PDF Monitoring
IP SCTP CHUNK Handover protocol data IP TCP/ UDP Passenger IP SCTP CHUNK IP TCP/ UDP Passenger data Effective 40,0 L2 Throughput in a WLAN 35,0 30,0 Throughput (Mbps) 25,0 20,0 5,0 0,0 5,0 0,0 msctp MIP 0 500 000 500 IP IP TCP/ UDP Passenger data Passenger IP packet size (bytes) IP IP TCP/ UDP Passenger data
IP SCTP CHUNK Handover protocol data IP TCP/ UDP Passenger CHUNK Passenger IP data TCP/ UDP Effective 40,0 L2 Throughput in a WLAN 35,0 30,0 Throughput (Mbps) 25,0 20,0 5,0 0,0 5,0 0,0 msctp MIP 0 500 000 500 IP IP TCP/ UDP Passenger data Passenger IP packet size (bytes) IP IP TCP/ UDP Passenger data
IBBT Tr@ins project Technical Business model Conclusions
Flowchart business case Forecasting Forecasting the number of train passengers Technology assignment Cost/benefit model Evaluation Determining rollout scheme Calculating potential Internet on train users Determining min. technology requirements Computing technical scenarios Calculating cost / benefit model Output business model
Forecasting Passenger forecasting 20% 8% 6% First class adoption Second class adoption 00000 4% Number of passeng gers per day 80000 60000 40000 Relation Oostende - Eupen 2% 0% 8% 6% 4% 2% 0% 20000 2008 2009 200 20 202 203 204 205 206 207 0 Adoption
Technology assignment
Technology mapping None - UMTS HSDPA WiMAX 2008 2009 200 20 202 UMTS + WiMAX
Cost parameters Wireless data networks Mobile networks Satellite networks CapEx Train equipment Outdoor antenna Indoor network Network equipment Trackside network NOC + + +++ +++ + + General costs Helpdesk Sales (billing) ++ ++ ++ OpEx Operations Marketing Maintenance and repair Network management +++ + ++ License costs Network connection Outdoor link + +++ ++
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) Service of 30 kbps: Prepaid: 3.4 / Subscription: 8 Two revenue scenarios Both first and second class paying First class free Internet, second class paying Modal switch from st to 2nd class!!!
NPV analysis (0 years, 5% discount rate, st class free Internet) 5 NP PV [M ] 0 5 0-5 -0 UMTS WIMAX UMTS + WIMAX WIMAX + UMTS WiMAX STAT. SAT WAY SAT 2 WAY -5 5 0 5 20 25 30 35 40 45 50 55 Bandwidth per user [kbps] 30 kbps per individual user user experience of ± Mbps
NPV analysis (0 years, 5% discount rate, st class free Internet) 5 NP PV [M ] 0 5 0-5 -0 UMTS WIMAX UMTS + WIMAX WIMAX + UMTS WiMAX STAT. SAT WAY SAT 2 WAY -5 5 0 5 20 25 30 35 40 45 50 55 Bandwidth per user [kbps] 30 kbps per individual user user experience of ± Mbps
NPV analysis (0 years, 5% discount rate, st class free Internet) 5 NP PV [M ] 0 5 0-5 -0 UMTS WIMAX UMTS + WIMAX WIMAX + UMTS WiMAX STAT. SAT WAY SAT 2 WAY -5 5 0 5 20 25 30 35 40 45 50 55 Bandwidth per user [kbps] 30 kbps per individual user user experience of ± Mbps
NPV analysis (0 years, 5% discount rate, st class free Internet) 5 NP PV [M ] 0 5 0-5 -0 UMTS WIMAX UMTS + WIMAX WIMAX + UMTS WiMAX STAT. SAT WAY SAT 2 WAY -5 5 0 5 20 25 30 35 40 45 50 55 Bandwidth per user [kbps] 30 kbps per individual user user experience of ± Mbps
NPV analysis (0 years, 5% discount rate, st class free Internet) 5 NP PV [M ] 0 5 0-5 -0 UMTS WIMAX UMTS + WIMAX WIMAX + UMTS WiMAX STAT. SAT WAY SAT 2 WAY -5 5 0 5 20 25 30 35 40 45 50 55 Bandwidth per user [kbps] 30 kbps per individual user user experience of ± Mbps
Sensitivity results UMTS CASE -00% -50% 0% 50% 00% Bandwidth per user -94% Number of SIM cards per train 2% Second class adoption % Mobile networks Due to the high bandwidth cost for UMTS Modal switch % Cellular bandwidth cost -% Full WiMAX network More users + more bandwidth (less important with dedicated networks) = higher revenue CASE WiMAX 5-00% -50% 0% 50% 00% Second class adoption 5% Bandwidth per user 30% Tariff per hour -0% Modal switch 4% Leasing percentage poles 2%
IBBT Tr@ins project Technical Business model Conclusions
Summary When more bandwidth is guaranteed to the customer: UMTS cases Only UMTS: Problems No QoS 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 Way 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 Influenced by many factors User forecasting, rollout scheme, offered bandwidth Business models Role and participation of the partners important
Thank you! Contacts: jan.vanooteghem@intec.ugent.be daan.pareit@intec.ugent.be