Dynamic Resource Scheduling Approaches in Server Less Computing

Abstract

Serverless computing has emerged as a transformative paradigm in cloud computing, offering event-driven execution and automated resource management without the need for explicit infrastructure provisioning. However, its dynamic, multi-tenant, and stateless nature introduces significant challenges in resource scheduling, particularly in maintaining a balance between performance, cost efficiency, and service-level agreements (SLAs). This paper presents a comprehensive review of dynamic resource scheduling approaches in serverless architectures, categorizing them into machine learning-based, heuristic, and resource-aware strategies. We analyse the strengths and limitations of each approach and discuss their applicability in heterogeneous and resource-constrained environments. Furthermore, the paper explores the role of serverless-aware orchestration tools and frameworks, including Kubernetes-based solutions, in enabling scalable and efficient function deployment. Finally, we identify open research challenges and propose future directions, including edge-serverless integration, sustainable scheduling, and AI-driven optimization for next-generation cloud-native systems.