Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Runtime Verification Containers for Publish/Subscribe Networks

Ali Mehran, Dogan Ulus

TL;DR

This work tackles the challenge of validating dynamic publish/subscribe networks at scale by proposing containerized runtime verification (RV) containers that operate as both subscribers and publishers to monitor, analyze, and report property violations in real time. It presents an end-to-end framework around the Reelay RV library, including message synchronizers and DDS↔Zenoh bridges, and benchmarks baseline pub/sub performance alongside RV overhead across multiple container networking configurations. A real-world Autoware/Carla case study demonstrates practical deployment on single and multi-host setups, highlighting the benefits and tradeoffs of containerized RV in software-defined vehicle scenarios. The results indicate that containerized RV can achieve high throughput with modest overhead, especially with Zenoh, and point to scalable architectural strategies (multi-property containers, pull-based observers, and binary wire formats) for industrial adoption.

Abstract

Publish/subscribe (pub/sub) networks are a cornerstone of modern distributed systems, playing a crucial role in applications like the Internet of Things (IoT) and robotics. While runtime verification techniques seem ideal for ensuring the correctness of such highly dynamic and large-scale networks, integrating runtime monitors seamlessly into real-world industrial use cases presents significant challenges. This paper studies modern containerization technology to deploy runtime verification tools to monitor publish/subscribe networks with a performance focus. Runtime verification containers are lightweight and deployable alongside other containerized publisher and subscriber participants. Each runtime verification container monitors message flow, enabling runtime verification of network behavior. We comprehensively benchmark the container-based approach using several experiments and a real-world case study from the software-defined vehicle domain.

Runtime Verification Containers for Publish/Subscribe Networks

TL;DR

This work tackles the challenge of validating dynamic publish/subscribe networks at scale by proposing containerized runtime verification (RV) containers that operate as both subscribers and publishers to monitor, analyze, and report property violations in real time. It presents an end-to-end framework around the Reelay RV library, including message synchronizers and DDS↔Zenoh bridges, and benchmarks baseline pub/sub performance alongside RV overhead across multiple container networking configurations. A real-world Autoware/Carla case study demonstrates practical deployment on single and multi-host setups, highlighting the benefits and tradeoffs of containerized RV in software-defined vehicle scenarios. The results indicate that containerized RV can achieve high throughput with modest overhead, especially with Zenoh, and point to scalable architectural strategies (multi-property containers, pull-based observers, and binary wire formats) for industrial adoption.

Abstract

Publish/subscribe (pub/sub) networks are a cornerstone of modern distributed systems, playing a crucial role in applications like the Internet of Things (IoT) and robotics. While runtime verification techniques seem ideal for ensuring the correctness of such highly dynamic and large-scale networks, integrating runtime monitors seamlessly into real-world industrial use cases presents significant challenges. This paper studies modern containerization technology to deploy runtime verification tools to monitor publish/subscribe networks with a performance focus. Runtime verification containers are lightweight and deployable alongside other containerized publisher and subscriber participants. Each runtime verification container monitors message flow, enabling runtime verification of network behavior. We comprehensively benchmark the container-based approach using several experiments and a real-world case study from the software-defined vehicle domain.
Paper Structure (19 sections, 6 figures, 2 tables)

This paper contains 19 sections, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Publish/Subscribe Networks
  4. Data Distribution Service (DDS)
  5. Eclipse Zenoh
  6. Modern Container Technology
  7. Related Work
  8. Containerization
  9. Target Architectures
  10. Container Networking Modes
  11. Container Orchestration
  12. Multi-stage Container Build
  13. Containerized Network Performance Evaluation
  14. Baseline Experiments in Containers
  15. Runtime Verification Experiments in Containers
  16. ...and 4 more sections

Figures (6)

  • Figure 1: Runtime verification containers on a publish/subscribe network
  • Figure 2: Container network architectures considered in the paper
  • Figure 3: Baseline network performance for DDS and Zenoh pub/sub protocols
  • Figure 4: Latency performance for containerized runtime verification
  • Figure 5: Container deployment setup using a single host machine
  • ...and 1 more figures