Green-Stream: Enabling Sustainable Decision- Making through Adaptive Computation Handover in Cloud-Edge Ecosystems

Abstract

Cloud-edge ecosystems offer a promising approach for real-time data processing, but require efficient resource allocation to achieve sustainability. This research addresses this challenge by introducing Green-Stream, a novel framework for adaptive computation handover in cloud-edge ecosystems. Green-Stream leverages machine learning to analyze real-time data on resource availability, energy consumption, and task complexity. Based on this analysis, it dynamically allocates tasks between the cloud and the edge, prioritizing energy efficiency while maintaining low latency for real-time decision-making. We evaluate Green-Stream on a dataset simulating a cloud-edge ecosystem with various resource profiles and task types. Performance metrics include energy consumption and task execution time. Compared to static allocation approaches, Green-Stream achieves significant reductions in energy consumption (average 20% decrease) while maintaining low latency for real-time tasks. This research paves the way for sustainable cloud-edge applications by enabling efficient resource utilization and minimized environmental impact.