Використання контейнерних застосунків для задачі прогнозування

Abstract

This work discusses the increasing need for effective big data analysis methods like forecasting and the challenge of inefficient resource management. Containerization technologies, such as Kubemetes, offer a solution to these problems by providing improved resource utilization, simplified deployment, and greater flexibility. Containerization allows for autoscaling, which adjusts resource allocation based on workload, improving performance, reducing costs, and enhancing forecasting accuracy. By combining autoscaling capabilities with Kubemetes' container orchestration, forecasting tasks can achieve better performance, flexibility, and cost-effectiveness, leading to widespread adoption in big data analysis.