Tetris: sn SLA-aware application placement strategy in the Edge-Cloud continuum.

Abstract

An Edge-Cloud Continuum integrates edge and cloud resources to deliver a flexible and scalable infrastructure. While this paradigm reduces latency and enhances scalability, the heterogeneous nature of these environments introduces challenges such as resource fragmentation and inefficient placement. Many existing approaches assume high resource availability and homogeneous infrastructures, overlooking fragmentation, application drops, and the interplay between performance metrics. This work introduces Tetris, an edge-cloud continuum application placement strategy inspired by the iconic puzzle game, designed to operate in a heterogeneous and complex edge–cloud continuum environment. It prioritizes tasks based on urgency, workload diversity, and resource availability. Compared to state-of-the-art approaches, Tetris reduces latency SLA violations by approximately 75%, while maintaining zero drop occurrences. Unlike proximity-based methods that focus on placing tasks near users, Tetris’s key strength lies in avoiding resource fragmentation, simplifying the placement problem, and improving overall system resilience. Additionally, clustering analysis reveals strong correlations between fragmentation and service degradation, reinforcing the importance of balanced resource allocation. These findings contribute to the development of more resilient, efficient, and explainable edge–cloud systems, with improved QoS and QoE for end-users.