A Profile-Based Approach to Just-in-Time Scalability for Cloud Applications

TLDR

Cloud platforms provide on‑demand resource utilization that enables runtime scaling, yet just‑in‑time scalability cannot be achieved by deployment alone and existing methods force developers to rewrite applications, binding them to specific infrastructure. This study proposes using profiles to capture expert knowledge for scaling diverse applications. The profile‑based approach automates deployment and scaling in the cloud and is demonstrated through a real‑world case study, showing its feasibility. Just‑in‑time scalability is achieved without tying applications to a particular cloud infrastructure.

Abstract

Cloud platforms offer resource utilization as on demand service, which lays the foundation for applications to scale during runtime. However, just-in-time scalability is not achieved by simply deploying applications to cloud platforms. Existing approaches require developers to rewrite their applications to leverage the on-demand resource utilization, thus bind applications to specific cloud infrastructure. In this paper, profiles are used to capture expertspsila knowledge of scaling different types of applications. The profile-based approach automates the deployment and scaling of applications in cloud. Just-in-time scalability is achieved without binding to specific cloud infrastructure. A real case is used to demonstrate the process and feasibility of this profile-based approach.