Scholarlink

<p>In recent years, the&#xa0;Internet of Things (IoT) technology has led to a growing acceptance of innovative IoT applications such as smart transportation and self-driving cars, smart networks, healthcare services, and immediate reactions that are highly sensitive to delay and demanding acceptable level of quality of service (QoS). The system's incoming workload fluctuating and the necessity to comply with resource providers' service level agreements necessitate the implementation of intelligent, autonomous, integrated, and flexible scaling mechanisms for edge servers, which missing in most researches. Due to the changing nature of resources over the time, static policies may not be suitable for such stochastic and dynamic environments. In this paper, we introduce a Joint Computation Offloading and Autonomous Resource Scaling solution utilizing reinforcement learning approach to combine computation offloading and edge resources auto-scaling focusing on offloading computations to local edge servers. Also, two modules, offloading and scaling managers, are embedded in the proposed solution architecture. They autonomously control the scaling of the available resources based on the parameters of remaining energy, latency, and cost, simultaneously making a decision to offload computations. Simulation results indicate that the proposed solution reduces cost by 8.9% and latency by 25.7% while increasing remaining energy by 0.9% compared to the other methods.</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

An efficient task offloading and auto-scaling approach for IoT applications in edge computing environment

  • 期刊论文

  • 发表日期：

    2025年5月2日

查看全文
  • Milad Ghahari-Bidgoli,
  • Mostafa Ghobaei-Arani,
  • Ahmad Sharif

摘要

In recent years, the Internet of Things (IoT) technology has led to a growing acceptance of innovative IoT applications such as smart transportation and self-driving cars, smart networks, healthcare services, and immediate reactions that are highly sensitive to delay and demanding acceptable level of quality of service (QoS). The system's incoming workload fluctuating and the necessity to comply with resource providers' service level agreements necessitate the implementation of intelligent, autonomous, integrated, and flexible scaling mechanisms for edge servers, which missing in most researches. Due to the changing nature of resources over the time, static policies may not be suitable for such stochastic and dynamic environments. In this paper, we introduce a Joint Computation Offloading and Autonomous Resource Scaling solution utilizing reinforcement learning approach to combine computation offloading and edge resources auto-scaling focusing on offloading computations to local edge servers. Also, two modules, offloading and scaling managers, are embedded in the proposed solution architecture. They autonomously control the scaling of the available resources based on the parameters of remaining energy, latency, and cost, simultaneously making a decision to offload computations. Simulation results indicate that the proposed solution reduces cost by 8.9% and latency by 25.7% while increasing remaining energy by 0.9% compared to the other methods.