EISim: A Platform for Simulating Intelligent Edge Orchestration Solutions

TL;DR

EISim is a new simulation platform built on PureEdgeSim that enables researchers to evaluate intelligent edge orchestration solutions, especially those based on deep reinforcement learning, in various scenarios involving task offloading and resource management.

Contribution

The paper introduces EISim, a novel extension of PureEdgeSim, specifically designed to support research on intelligent edge orchestration methods with new tools and capabilities.

Findings

Supports simulation of deep reinforcement learning solutions

Enables testing of different orchestration control topologies

Includes tools for environment creation, training, and visualization

Abstract

To support the stringent requirements of the future intelligent and interactive applications, intelligence needs to become an essential part of the resource management in the edge environment. Developing intelligent orchestration solutions is a challenging and arduous task, where the evaluation and comparison of the proposed solution is a focal point. Simulation is commonly used to evaluate and compare proposed solutions. However, the currently existing, openly available simulators are lacking in terms of supporting the research on intelligent edge orchestration methods. To address this need, this article presents a simulation platform called Edge Intelligence Simulator (EISim), the purpose of which is to facilitate the research on intelligent edge orchestration solutions. EISim is extended from an existing fog simulator called PureEdgeSim. In its current form, EISim supports simulating deep reinforcement learning based solutions and different orchestration control topologies in scenarios related to task offloading and resource pricing on edge. The platform also includes additional tools for creating simulation environments, running simulations for agent training and evaluation, and plotting results.