Enhancing Autoscaling in Cloud Environment: Case of Hotels Network Management System
Kvitoslava Obelovska, Orest Mamchuk, Андрій АБЗЯТОВ, Michal Greguš, Oleh Liskevych, Rostyslav Liskevych · Procedia Computer Science · 2024
This study aims to improve the provision of services in a multi-client oriented system, the integrated incoming flow of requests from which has a dynamic, unpredictable nature of varying intensity and distribution. To serve such flows when using cloud environments, automatic scaling is typically recommended. However, in some applications, these methods require improvements in terms of response time and system resilience in case of sudden changes in input flows. In this work, a model of automatic scaling is proposed, which involves the collection and special processing of additional metrics. These metrics measure the actual state of incoming integrated requests from multiple customers. Based on metrics, at a certain point in time an alarm is raised, which causes the activation of an additional server or its removal. Considering additional indicators, it is possible to update the network infrastructure more flexibly. In the developed implementation, the hotel business is chosen as an example, but the proposed approach can be used for applications with similar requirements in different areas of human activity.