Fig 1.
Fig 2.
Typical architecture of Kafka [6].
Table 1.
Summary of reviewed data sources and communication link using in TMS.
Fig 3.
Proposed architecture for road traffic morning application.
Fig 4.
PhayaThai-1 unreachable case.
Fig 5.
PhayaThai-6 unreachable case.
Fig 6.
Operational flow chart of Kafka producer application with local broker switching mechanism inside the mesh node.
Fig 7.
Operational flow chart of Kafka consumer application in local traffic controller box.
Fig 8.
Operational flow chart of traffic controller box application to relay locally received messages to external Kafka broker.
Fig 9.
Operational flow chart of external Kafka consumer application in central traffic command centre.
Fig 10.
Location of local broker 1 at the junction of Phaya Thai road and Rama 4 road (near Sam Yan).
Fig 11.
Location of local broker 2 at the junction of Phaya Thai Road and Chulalongkorn Soi 12 (near Mahboonkrong).
Fig 12.
Installation of local broker 1 at specific traffic police controller box in (a) and (b).
Fig 13.
Installation of local broker 2 at specific traffic police controller box.
Fig 14.
Components for wireless mesh node inside waterproof box.
Fig 15.
Snapshot example of monitoring application at external broker.
Fig 16.
End-to-end average latency of image messages received at each broker for each wireless node (Phaya Thai road test).
Fig 17.
Image message loss percentage measurement by traffic monitoring application at local brokers and external broker (Phaya Thai road test).
Fig 18.
Local broker switching time of PhayaThai-2 and PhayaThai-5 in seconds (Phaya Thai road test).
Fig 19.
Practical operation status of SDWMN control plane for 13 hours testing (Phaya Thai road test).
Fig 20.
Laboratory-based emulated outdoor network.
Fig 21.
Routing of OLSR network.
Fig 22.
Routing of SDWMN network.
Table 2.
Different cases tested for the comparison.
Fig 23.
Example of case 1 for OLSR and SDWMN.
Fig 24.
Example of case 2 for OLSR and SDWMN.
Fig 25.
Example of case 3 for OLSR and SDWMN.
Fig 26.
Time needed to re-route to alternative route.
Fig 27.
Time needed to restore to primary route.
Fig 28.
End-to-end average latency for PhayaThai-2 node for primary route and alternative route after re-routing/restoration completion.