Impact of EM a,ack signatures on ETCS Quality of Service Indicators. Marina AGUADO, UPV/EHU ChrisHan PINEDO, UPV/EHU Eduardo JACOB, UPV/EHU

Transcription

1 Impact of EM a,ack signatures on ETCS Quality of Serice Indicators Marina AGUADO, UPV/EHU ChrisHan PINEDO, UPV/EHU Eduardo JACOB, UPV/EHU

2 Goal EM A%ack signatures on Communica4on Systems Assess the impact of the jammers on the applica4on leel and not only on the physical layer. GOAL Interferences in the CommunicaHon Architecture Railway Network Aailability ApplicaHon considered: ETCS communicahon between the train and the RBC for ERTMS leel 2 and leel 3. 2

3 Goal (cont.) Perform simula4ons and correlate results in order to iden4fy network measures (packet loss, delay, ) that could be used as an addi4onal sensor to detect abnormal behaiour or a%acks. Failure or Jammers Characteriza4on INPUT Railway Integrated SimulaHon Framework OUTPUT EIRENE or QoS KPI s Packet error rate ConnecHon Establishment End to end delay Health Attack Manager 3

4 ERTMS ETCS Application ETCS application protocol ETCS Application Open Communication system GSM-R Radio protocol Open Communication system GSM-R Train-Side GSM-R radio transmission Track-Side ERTMS = FuncHonal Subsystem + Telecom Subsystem ** CHALLENGE = Modelling both Subsystems (FuncHonal Subsystem + Telecom Subsystem ) in the same simulahon integrated framework

5 SimulaHon Framework

6 SimulaHon Framework Interferred Scenario Analysis in Simulation Platform Railway Network Informa4on (Train Program) Number of trains RBC Balise locahon Terrain InformaHon n Communica4on Network Topology PropagaHon Model + Failure or Jammers Characteriza4on Demo Presenta4on: w w w INPUT Scenario 1 : Legacy architecture Scenario 2 : Next Genera4on IP LTE with failures Scenario 3 Conges4on Railway Integrated SimulaHon Framework OUTPUT EIRENE or QoS KPI s Packet error rate ConnecHon Establishment Delay End to end delay Unnecessary Emergency Brakes Aailability Health Attack Manager Further detail on SimulaHon Framework internals: Modelling and simulahon of ERTMS for current and future mobile technologies. Interna4onal Journal of Vehicular Technology (2015) Towards zero on-site teshng: Adanced traffic management & control systems simulahon framework including communicahon KPIs and response to failure eents. In Wireless Vehicular Communica4ons (WiVeC) 2014 IEEE Vehicular Technology Communica4on Conference

7 SimulaHon Framework 7

8 SimulaHon Scenario Core-Network Main funchon: Forward packets between the RBC and antennas. Modelled the normal condi4ons for a end-toend (train-to-ground) link 10 m Access-Network Main funchon: Forward packets between wireless and wired interfaces Equialent to a BTS of GSM-R. Transmission Power set to proide between -95dBm and -93dBm at 3 km according to field measures (4,27mW). 6 km 102 km 8

9 SimulaHon Scenario Detail on ETCS deices Modelled ETCS (ATP) applica4on and lower layer protocols 9

10 Real Grey Jammer Real Siler Jammer SimulaHon Framework: Jammers Modelled as a jammer with a constant output power of -10 dbm in the MHz frequency band. Thus, it affects uplink and downlink GSM-R channels. Modelled as a single band jammer of 4MHz centred in the 923 MHz frequency with a cyclic transmission power of 8usec between -5 dbm and 7,5 dbm. Thus, it only affects downlink GSM-R channels. Modelled Siler and Grey Jammers 10

11 Test Methodology Jammered Scenario Parameterize network Focus geographical scope to two BTSs (6 km) Baseline Scenario Select the jammer. Two jammers modelled Locate the Jammer Middle of two BTSs Close to one BTS In the same locahon of one BTS Eery 25 meters between the two BTSs: Locate the train Perform mulhple simulahons to obtain stahshcally alid results Thousands of simula4ons per Baseline Scenario and Jammered Scenario!! 11

12 SimulaHon Results for Siler Jammer (II) Although the jammer in the noisy area is not so powerful to block totally the signal, it increases the packet losses and so the end-to-end delay (due to packet retransmissions). Once 5% of packet loss is passed, e2e delay increases significantly A connected link doesn t imply a alid link for ETCS. 12

13 1. Our simula4on framework allows assessing the impact of jammering deices in the applica4on layers (i.e., ETCS). Easily ealuate mulhple scenarios with a broad set of parameters (layout, type, locahon and power of the jammer, ). IdenHfy the dependency between KPIs and those parameters. IdenHfy weak points and proide recommendahons (oerlaping of coerage areas, ). 2. Simula4on results proe that informa4on proided by the network deices could be used as an addi4onal sensor to detect abnormal behaiour or a%acks. Conclusion s For example, SNR cures might help to detect the presence of jammering deices. 13

14 Impact of EM a,ack signatures on ETCS Quality of Serice Indicators Marina AGUADO, UPV/EHU ChrisHan PINEDO, UPV/EHU Eduardo JACOB, UPV/EHU