A Docker Container-Based Dynamic Allocation System for Computational Resources

As data scale and computation demand increase, traditional centralized computing are facing several challenges. Distributed computing and edge computing in particular handle those problems by extending computation to the network edge, allowing faster data processing and instant responses. A centralized platform is important for efficient task scheduling, resource allocation and system stability. In this study, we propose a Docker container-based dynamic resource allocation system designed to communicate between server and multiple clients. The system enables task transmission and real-time resource monitoring across clients. A dynamic allocation algorithm is applied to computation tasks such as model training and inference. The performance is evaluated by three metrics. Experiments show that the algorithm significantly reduces task latency, improves resource utilization, and achieves a more balanced load distribution. The findings highlight the potential for further development and application of dynamic resource allocation algorithms in computation platforms.