Performance Evaluation of the Nature-Inspired Algorithms for Container Scheduling for Elastic Containerized Multi-Cloud

In the dynamic world of elastic containerized multi-cloud environments, task scheduling algorithms are pivotal in maximizing resource utilization to attain performance objectives. This research focuses on several container scheduling algorithms based on Ant Colony Optimization, Chicken Swarm Optimization, Genetic Algorithm, Bee Colony Optimization, and Particle Swarm Optimization. In this respect, it seeks to evaluate these algorithms over metrics such as makespan, reliability, idle time, energy utilization, response time, CPU utilization, execution cost, and task migration. The paper analyzes these factors to identify the strengths and weaknesses of each algorithm. Using the standardized configuration of 50 containers and 5000 tasks, the evaluation gives the following results: all the algorithms have excellent reliability values of 0.99, a very good indication of their efficiency in the execution of tasks with minimal failure cases. ACO has an excellent CPU utilization of 86% and the lowest make span of 10 ms, indicating its efficiency in allocating resources and executing tasks. BCO has the highest CPU utilization at 88% and energy utilization at 1900 Joules. This indicates that it can effectively and very well exploit resources at the cost of increased energy consumption. ACO has the lowest response time of 7 ms, which ensures it responds quickly in general, and also with the lowest execution cost of $110, which indicates it is cost-effective. The performance of CSO and PSO is somewhat close to each other; GA has a balanced performance overall. To implement the abovementioned approach Container Cloud sim 4.0.0 is used. The paper presents a comparative study that helps to get deep insights into the performance characteristics of the different container scheduling algorithms for making an appropriate decision regarding deploying an application in elastic containerized multi-cloud environments.