clock synchronization in omnet
play

Clock Synchronization in OMNeT++ Henning Puttnies, Peter Danielis, - PowerPoint PPT Presentation

Institute of Applied Microelectronics and Computer Engineering A Simulation Model of IEEE 802.1AS gPTP for Clock Synchronization in OMNeT++ Henning Puttnies, Peter Danielis, Enkhtuvshin Janchivnyambuu, Dirk Timmermann University of Rostock,


  1. Institute of Applied Microelectronics and Computer Engineering A Simulation Model of IEEE 802.1AS gPTP for Clock Synchronization in OMNeT++ Henning Puttnies, Peter Danielis, Enkhtuvshin Janchivnyambuu, Dirk Timmermann University of Rostock, Germany

  2. Institute of Applied Microelectronics and Computer Engineering 1. Motivation  Real-time Ethernet systems No open standard established  Only proprietary solutions (expensive)   A standard-based approach is required  IEEE 802.1 Time-Sensitive Networking (TSN) Task Group  gPTP is a part of TSN standards (for sync) 10/09/2018 OMNeT++ Community Summit 2018 2

  3. Institute of Applied Microelectronics and Computer Engineering 2. Basics  Overview of gPTP protocol End station Bridge End station Types of time-aware systems  - End stations, bridges Types of ports  - Master, slave, passive Time-aware systems only communicate gPTP information directly  with other time-aware systems  Hop by hop synchronization 10/09/2018 OMNeT++ Community Summit 2018 3

  4. Institute of Applied Microelectronics and Computer Engineering 2. Basics  Best master clock selection (BMCS) Announce Announce Grandmaster Slave M End station M S Bridge S End station P All time-aware systems participate in BMCS  Announce message: time-synchronization spanning tree vector  Automatic changeover to a secondary grandmaster  M Master port S Slave port P Passive port 10/09/2018 OMNeT++ Community Summit 2018 4

  5. Institute of Applied Microelectronics and Computer Engineering 2. Basics  Propagation delay measurement End station 1 Bridge 1 Timestamps Delay requester Delay responder known by time time delay requestor t 1 t 1 t AB Pdelay_req f requestor r = t 2 f responder t 3 Pdelay_resp t BA t 2 t 1 t 2 t 4 t 4 Pdelay_resp_follow_up t 3 t 1 t 2 t 4 t 3 10/09/2018 OMNeT++ Community Summit 2018 5

  6. Institute of Applied Microelectronics and Computer Engineering 2. Basics  Propagation delay measurement  Rate ratio End station 1 End station 2 ( t 12 – t 11 ) r = Master Slave ( t 22 – t 21 ) port port t 11 r = 1 Clock es1 = Clock es2 Sync t 21 r < 1 Clock es1 < Clock es2 t 12 r > 1 Clock es1 > Clock es2 Sync t 22 ( t 12 – t 11 ) ≠ ( t 22 – t 21 ) 10/09/2018 OMNeT++ Community Summit 2018 6

  7. Institute of Applied Microelectronics and Computer Engineering 2. Basics: Transport of Sync. Information Bridge Grandmaster Slave Master Slave Slave Master port port port port t e11 Sync t e21 t b11 Follow_up correctionField(1) t b21 Compute correctionField(2) t b12 Sync t b22 Follow_up t b31 correctionField(2) t b41  correctionField: Composed of propagation delay and residence time  Slave: preciseOriginTimestamp + <delayToGM>  Synced to GM time 10/09/2018 OMNeT++ Community Summit 2018 7

  8. Institute of Applied Microelectronics and Computer Engineering 3. Implementation  Scope of the project gPTP simulation model in OMNeT++ using the INET library  Integrate gPTP model seamlessly with other protocols from INET  Implement only time synchronization and propagation delay measurement  Best master clock not part of project Assumption: GM shall no be selected randomly  Implement simple clock with constant drift 10/09/2018 OMNeT++ Community Summit 2018 8

  9. Institute of Applied Microelectronics and Computer Engineering 3. Implementation  Model of clock with constant drift C(t) = at + b b T(t) = t t [s] 10/09/2018 OMNeT++ Community Summit 2018 9

  10. Institute of Applied Microelectronics and Computer Engineering 3. Implementation  Model of clock with constant drift Bridge Grandmaster Slave Slave Master Master Master port port port port t b11 t e11 Sync t b12 Sync t b51 t e12 t b21 Follow_up Follow_up t b52 t b22 Drift calculation Drift calculation when Sync and when Sync and Follow_up are sent t b31 Follow_up are received Sync t b32 t b61 t e21 Follow_up Sync t b62 t e22 t b41 Follow_up t b42 10/09/2018 OMNeT++ Community Summit 2018 10

  11. Institute of Applied Microelectronics and Computer Engineering 3. Implementation  Model of gPTP functionalities Grandmaster Slave M End station M S Bridge S End station P • Eth. interface with gPTP support: EthernetInferfaceGPTP • Simple module for gPTP functions: etherGPTP 10/09/2018 OMNeT++ Community Summit 2018 11

  12. Institute of Applied Microelectronics and Computer Engineering 3. Implementation  Model of time-aware systems Grandmaster Slave M End station M S Bridge S End station P • Simple module tableGPTP • Simple module clock • Compound module EthernetInferfaceGPTP 10/09/2018 OMNeT++ Community Summit 2018 12

  13. Institute of Applied Microelectronics and Computer Engineering 4. Evaluation: Simulation Setup • Same setup as Lim et al.* (BMW + TUM) • Evaluation: Propagation delay measurement • Time difference to GM • (before resynchronization) Clock drift of time-aware systems in domain [ppm] Master Bridge0 Bridge1 Bridge2 Slave0 Slave1 Slave2 Slave3 Slave4 Slave5 Slave6 Slave7 0 30 -15 20 -50 10 50 -5 -50 40 -15 -35 *Hyung-Taek Lim, Daniel Herrscher and Lars Volker “IEEE 802.1AS Time Synchronization in a switched Ethernet based In- Car Network”, IEEE VNC 2011 10/09/2018 OMNeT++ Community Summit 2018 13

  14. Institute of Applied Microelectronics and Computer Engineering 4. Evaluation: Propagation Delay Measurement • Converge to 25 ns (absolute difference < 0.5 ns) • Lim et al.: +/- 10 ns acceptable Propagation Absolute Node Error (%) delay[ns] difference [ns] Slave 0 25.43 1.72% 0.43 Slave 1 25.43 1.72% 0.43 Slave 2 24.78 -0.88% 0.22 Slave 3 25.29 1.16% 0.29 Slave 4 25.29 1.16% 0.29 Slave 5 24.78 -0.88% 0.22 Slave 6 24.78 -0.88% 0.22 Slave 7 25.43 1.72% 0.43 Bridge 0 25 0.00% 0.00 Bridge 1 25.43 1.72% 0.43 Bridge 2 24.78 -0.88% 0.22 10/09/2018 OMNeT++ Community Summit 2018 14

  15. Institute of Applied Microelectronics and Computer Engineering 4. Evaluation: Time Difference to GM Time difference to GM (before resynchronization)  As expected: e.g., for 125ms and +/- 50ppm  +/- 6.25us  Time difference to GM before resynchronization in our implementation [µs] Node Sync interval Sync interval 62.5 ms 125 ms 2.36 4.24 Bridge 0 -3.12 -6.25 Slave 0 0.63 1.25 Slave 1 -2.19 -4.37 Slave 7 -0.94 -1.87 Bridge 1 3.13 6.25 Slave 2 -0.94 -1.87 Slave 6 2.50 5.00 Slave 5 1.25 2.50 Bridge 2 -0.31 -0.63 Slave 3 -3.12 -6.25 Slave 4 10/09/2018 OMNeT++ Community Summit 2018 15

  16. Institute of Applied Microelectronics and Computer Engineering 5. Conclusion  We have contribute  Simulation model of gPTP  Models for time-aware systems: end-station and bridge  Simple clock model with constant drift  Comparisons of results to literature  Useful in simulating any networks based on the gPTP  Entire system is publicly available*  Future work: Utilize other the clock models * https://gitlab.amd.e-technik.uni-rostock.de/peter.danielis/gptp-implementation 10/09/2018 OMNeT++ Community Summit 2018 16

  17. Institute of Applied Microelectronics and Computer Engineering Thank you for your attention. Questions? 10/09/2018 OMNeT++ Community Summit 2018 17

Recommend


More recommend