Scholarlink

<p>Vehicular Edge Computing (VEC) is a decentralized paradigm for processing and managing the vast amount of data generated by connected vehicles and roadside IoT devices. Vehicular Edge Nodes (VENs), such as roadside units (RSUs), onboard vehicular gateways, or micro-data centers deployed near roadways, serve as edge nodes that host and execute vehicular services and applications. Microservices offer promising deployment strategies for future VEC applications such as autonomous driving, vehicular platoons, infotainment, etc. However, a critical challenge in VEC lies in selecting appropriate edge nodes for deploying these microservices, given that the edge nodes are spatially distributed across the vehicular environment and possess constrained computational and storage capacities. These microservices are dynamically deployed on the VENs in order to minimize service latency and optimize resource consumption. However, the high mobility of vehicles and the limited computational capacity of VENs pose significant challenges in maintaining efficient service delivery. To address these issues,we propose a comprehensive microservice placement framework that caters to the mobility of vehicles and facilitates service migration across VENs. Specifically, we formulate the placement decisions as a multi-objective optimization problem, considering latency constraints,VEN resource limitations, vehicular mobility, and service migration among VENs. Our model employs a KDTree-based nearest neighbor strategy for modelling the spatio-temporal dynamics of VENs. The model is solved using standard evolutionary algorithms such as NSGA-II, ACO, PSO, and DE. The framework supports adaptive migration and offloading to ensure seamless service continuity. We evaluate the proposed model using real-world vehicular mobility traces from Luxembourg City. The results show that the best-performing optimizer depends on the target objective: DE achieves the lowest service latency, PSO minimizes resource consumption, and NSGA-II reduces migration and offloading while maximizing edge execution, highlighting the importance of multi-objective placement strategies in dynamic vehicular edge environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mobility aware microservice placement in vehicular edge computing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月9日

查看全文
  • A. Surayya,
  • Md Muzakkir Hussain,
  • Anwar Ulla Khan,
  • Leila Jamel,
  • Ateyah Alzahrani

摘要

Vehicular Edge Computing (VEC) is a decentralized paradigm for processing and managing the vast amount of data generated by connected vehicles and roadside IoT devices. Vehicular Edge Nodes (VENs), such as roadside units (RSUs), onboard vehicular gateways, or micro-data centers deployed near roadways, serve as edge nodes that host and execute vehicular services and applications. Microservices offer promising deployment strategies for future VEC applications such as autonomous driving, vehicular platoons, infotainment, etc. However, a critical challenge in VEC lies in selecting appropriate edge nodes for deploying these microservices, given that the edge nodes are spatially distributed across the vehicular environment and possess constrained computational and storage capacities. These microservices are dynamically deployed on the VENs in order to minimize service latency and optimize resource consumption. However, the high mobility of vehicles and the limited computational capacity of VENs pose significant challenges in maintaining efficient service delivery. To address these issues,we propose a comprehensive microservice placement framework that caters to the mobility of vehicles and facilitates service migration across VENs. Specifically, we formulate the placement decisions as a multi-objective optimization problem, considering latency constraints,VEN resource limitations, vehicular mobility, and service migration among VENs. Our model employs a KDTree-based nearest neighbor strategy for modelling the spatio-temporal dynamics of VENs. The model is solved using standard evolutionary algorithms such as NSGA-II, ACO, PSO, and DE. The framework supports adaptive migration and offloading to ensure seamless service continuity. We evaluate the proposed model using real-world vehicular mobility traces from Luxembourg City. The results show that the best-performing optimizer depends on the target objective: DE achieves the lowest service latency, PSO minimizes resource consumption, and NSGA-II reduces migration and offloading while maximizing edge execution, highlighting the importance of multi-objective placement strategies in dynamic vehicular edge environments.