Handling scheduling uncertainties through traffic shaping in Time-Triggered train networks

  • Qinghan Yu
  • , Xibin Zhao
  • , Hai Wan
  • , Yue Gao
  • , Chenyang Lu
  • , Ming Gu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations

Abstract

While trains traditionally relied on field bus to support real-time control applications, next-generation trains are moving toward Ethernet as an integrated, high-bandwidth communication infrastructure for real-time control and best-effort consumer traffic. Time-Triggered Ethernet (TT-Ethernet) is a promising technology for train networks because of its capability to achieve deterministic latencies for real-time applications based on pre-computed transmission schedules. However, the deterministic scheduling approach of TT-Ethernet faces significant challenges in handling scheduling uncertainties caused by switch failures and legacy end devices in train networks. Due to the physical constraints on trains, train networks deal with switch failures by bypassing failed switches using a short circuiting mechanism. Unfortunately, this mechanism incurs scheduling errors as frames bypassing the failed switch may arrive ahead of the pre-computed schedule, resulting in early, unexpected, and out of order arrivals. Furthermore, as trains evolve from traditional communication technologies to TT-Ethernet, the network must support legacy end devices that may generate frames at times unknown to the TT-Ethernet. We propose a novel traffic shaping approach to deal with scheduling uncertainties in TT-Ethernet. The traffic shaper of a TT-Ethernet switch buffers early frames and then releases them at their pre-scheduled arrive time. Furthermore, we devise an efficient buffer management method for the traffic shaper in face of fault scenarios. Finally, we use the traffic shaper to integrate legacy devices into TT-Ethernet. We have implemented the traffic shaping approach in a 24-port TT-Ethernet switch specifically designed for train networks. Experiments show the traffic shaping strategy can effectively deal with scheduling uncertainties incurred by switch failures and legacy devices.

Original languageEnglish
Title of host publication2017 IEEE/ACM 25th International Symposium on Quality of Service, IWQoS 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509019830
DOIs
StatePublished - Jul 5 2017
Event25th IEEE/ACM International Symposium on Quality of Service, IWQoS 2017 - Vilanova i la Geltru, Spain
Duration: Jun 14 2017Jun 16 2017

Publication series

Name2017 IEEE/ACM 25th International Symposium on Quality of Service, IWQoS 2017

Conference

Conference25th IEEE/ACM International Symposium on Quality of Service, IWQoS 2017
Country/TerritorySpain
CityVilanova i la Geltru
Period06/14/1706/16/17

Fingerprint

Dive into the research topics of 'Handling scheduling uncertainties through traffic shaping in Time-Triggered train networks'. Together they form a unique fingerprint.

Cite this