Skip to main navigation Skip to search Skip to main content

Stability Evaluation and Enhancement in Switched Networks: Detailed Delay Jitter Analysis Through Flow Interference

  • Xinyue Zhang
  • , Xuan Zhou
  • , Ershuai Li
  • , Feng He*
  • *Corresponding author for this work
  • Beihang University

Research output: Contribution to journalArticlepeer-review

Abstract

The stability of flow transmission is critical in real-time switched networks with deterministic traffic, such as avionics Ethernet, industrial control systems, and Time-Sensitive Networking (TSN). Delay jitter, as a key indicator of stability, captures the variability in packet arrival times that can affect temporal predictability. Traditional methods relying solely on worst-case delay evaluations often neglect finer delay variations within each flow. To address this gap, we propose a novel approach based on delay jitter variance to more precisely capture detailed delay fluctuations. This is achieved through comprehensive interference analysis, identifying two primary contributors to delay variability: interference value and time interval. Together, these factors determine the magnitude of delay jitter and its distribution over time. The stability evaluation can then be refined from bounded delay jitter to detailed jitter variance, which also offers guidance for stability enhancement. Based on these insights, we develop a heuristic routing optimization algorithm to improve stability in large scale networks. Experimental results demonstrate significant improvements in network stability without sacrificing real-time performance.

Original languageEnglish
JournalIEEE Internet of Things Journal
DOIs
StateAccepted/In press - 2026

Keywords

  • Stability analysis
  • interference analysis
  • routing optimization
  • switched networks

Fingerprint

Dive into the research topics of 'Stability Evaluation and Enhancement in Switched Networks: Detailed Delay Jitter Analysis Through Flow Interference'. Together they form a unique fingerprint.

Cite this