Hemanth's Internet Home
Back to TILs

Container-based Platform for Edge-Computing in Energy System Applications

Date: 2026-06-10

link: https://ieeexplore.ieee.org/abstract/document/10688966

Motivation

Designing power systems with the advent of microgrids and IoT based solutions for management has grown increasingly complex both in terms of communication networks and power grid deployments using microgrids. The Authors propose a container-based platform for Edge computing emulation based on a co-simulation model to implement, test, and validate Edge computing-based energy system application using ns-3 and OPAL-RT.

Methodology

  • Separated the validation architecture onto two distinct computing systems, one being OPAL-RT, a dedicated real-time digital hardware simulator for physical execution, and a separate Linux computer for computer network emulation using ns-3.
  • Implemented the complete multi-microgrid power system framework, tie-converters, and lines on an OPAL-RT (OP4610XG) digital simulator to compute the physical system dynamics in real time.
  • Constructed the decentralized communication network layer on the Linux computer using isolated Docker containers configured to serve as localized controllers, running Python-based management agents.
  • Configured each edge container node with two virtual interfaces (veth1 and veth2); veth1 routes multi-agent communications through ns-3 via a Linux bridge, while veth2 transmits measurements and control actions to the OPAL-RT machine.
  • Routed all node-to-node inter-container data communications through the core libraries of ns-3, modeling customized data rates, packet sizes, and propagation delay variations.
  • Used shared-memory data buffering mechanism to mve time-sensitive measurement updates and actuation signals between the OPAL-RT simulator and the Docker containers.
  • Demonstrated platform viability via a consensus control protocol case study over a directed communication graph using five containerized microgrid agents to achieve cooperative power-sharing and voltage regulation.
  • Modeled network resilience by configuring an additional malicious node inside the ns-3 topology to launch a Denial-of-Service (DoS) packet-flooding attack against a targeted microgrid container controller.

Drawbacks

  • Routing edge communications through a layered architecture of dual virtual interfaces (veth), Linux host bridges, and tap connections have a lot of scheduling overheads for this purpose.
  • Changes made to the physical power grid model inside OPAL-RT do not map to ns-3 network layouts.
  • Very similar to all the previous implementations, hence comes with the same Drawbacks.