.png){kind=icon}
.png){kind=icon}
Return
Intent-Driven VM Allocation Strategy for Optimizing Cloudlet Processing in Edge–Cloud Computing
.png){kind=icon}
PRE
AI
DOI：10.1109/JIOT.2025.3639255.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Edge–cloud computing refers to a paradigm that combines the benefits of edge and cloud computing to optimize data processing and resource utilization. Edge–cloud computing plays a crucial role in resource allocation by optimizing the distribution of computational resources between edge devices and centralized cloud infrastructures. In the rapidly evolving landscape of edge–cloud computing, efficient virtual machine (VM) allocation is critical for optimizing resource utilization, minimizing latency, and ensuring high service-level agreement (SLA) compliance. This article introduces a novel heuristic VM allocation strategy, named least loaded and closest to deadline (LLCD), to enhance cloudlet or task processing in edge–cloud datacenters. By employing a heuristic approach inspired by mixed-integer nonlinear programming models, this strategy dynamically assigns VMs based on their current load and the impending deadlines of tasks, significantly reducing overall system latency and enhancing SLA success rates. Simulation was conducted across various computational intensities. The findings reveal that the proposed approach substantially improves resource utilization and operational efficiency, adapting to dynamic workloads, by achieving an SLA success ratio as 74.26% and 83.7% in different deadline scenarios. The adaptive nature of the LLCD algorithm allows real-time task reallocation based on system feedback, which mirrors the operational principles of AI-driven orchestration in distributed IoT environments. The validation is achieved through a multi-iteration simulation model that emulates dynamic IoT workloads, demonstrating LLCD’s learning capability in maintaining SLA stability and consistent latency reduction across changing task distributions. Moreover, the proposed heuristic provides a foundation for latency-efficient and learning-based management in distributed computing environments.
Keywords:
Cloud computing
edge computing
latency
makespan
propagation delay
service-level agreement (SLA)
Journal
IF:
8.9
Papers:
1.4W
Citations:
7.8W