Horizontal Scaling Implementation with Container Orchestrator. Reliability Analysis during Stress and Performance Testing

Autor: DINCA, A.-M., AXINTE, S.-D., BACIVAROV, I., PETRICA, G.
Jazyk: angličtina
Rok vydání: 2024
Předmět:
Zdroj: Advances in Electrical and Computer Engineering, Vol 24, Iss 3, Pp 23-32 (2024)
Druh dokumentu: article
ISSN: 1582-7445
1844-7600
DOI: 10.4316/AECE.2024.03003
Popis: Software platforms’ infrastructure encompasses the collection of physical resources used concomitantly for user interactions, computational or persistence operations. Virtualization facilitates dynamic allocation of resources to a platform, while staying within the infrastructure's limits, enabling deployments of multiple environments on a server, maximizing the computational power. This paper focuses on practical applications of virtualization principles on software platforms, and the analysis of performance indicators. Section II is centered on the Docker feature that packages the source code into images, facilitating testing in local containers, preparing them for the container orchestrator, Kubernetes that designs, creates and scales infrastructure components. The platform was scaled horizontally, because this method accommodates the application’s technical requirements, as stated in Section V. To identify the underlying causes of an inefficient auto-scaling infrastructure, a Fishbone analysis was conducted, followed by a novel proposal of triggering scaling operations, using AI tools that decide when to rescale an infrastructure, instead of traditional alert levels. The simple and advanced health check mechanisms are studied and implemented in Section VI, through separate APIs that validate the system’s status and dependencies automatically and periodically using Docker commands, proving the misleadingness of the container for the former, demonstrating the importance of such progressive mechanisms.
Databáze: Directory of Open Access Journals
Externí odkaz: