Інформаційна технологія автоматичної оптимізації мультирегіональної інфраструктури

The article presents an information technology aimed at the dynamic optimization of multi-region cloud deployments through the combined use of latency metrics, CDN analytics, and region-specific cloud pricing. The purpose of the work is to enhance the efficiency of cloud infrastructures by enabling automated region selection that reflects the actual geographical distribution of users and current cloud resource costs. The proposed solution includes a formal mathematical model in which effective latency is defined as a weighted minimum latency for user groups, while the total cost is calculated as the aggregated cost of active regions. An integral quality criterion is introduced to evaluate configuration alternatives. A software agent is developed to implement the proposed approach. The agent automatically collects real-time Cloudflare analytics, including request distribution and latency measurements, as well as cloud pricing data retrieved through provider APIs. Based on the formal model, the agent periodically evaluates the active deployment configuration, identifies more optimal alternatives, and initiates reconfiguration using Infrastructure-as-Code mechanisms. A quantitative case study is provided to demonstrate the practical applicability of the approach. The results show that combining European and South Asian regions in a two-region deployment may reduce effective latency by approximately 40 %, while annual cost savings of up to USD 2800 can be achieved in scenarios with favourable cost–latency trade-offs. The research confirms that the integration of operational latency telemetry with regional pricing information enables more efficient management of multi-region cloud systems compared to static region selection. The proposed information technology may serve as a foundation for further developments in adaptive cloud orchestration, including predictive traffic-aware scaling and integration with multi-cloud environments.