Collaborative offloading decision policy framework in IoT using edge computing

Archana Shirke

doi:10.1007/s11042-023-14383-4

2023-01-13

PRE

Archana Shirke *

Madhav Chandane

DOI：10.1007/s11042-023-14383-4

原文链接

原文求助

摘要

En 中文

Internet of Things (IoT) gives rise to concerns regarding edge computing policies for intelligent data processing to optimize resources at edge devices. The resources like energy, computation power, available memory, execution time need saving on for constraint-based IoT devices. These resources optimize to proper utilization of Edge devices, which increases the lifetime. A resource optimization decision is the basis of offloading some tasks from edge devices to the next level gateway/ server devices. This decision of full, partial, or no offloading depends on the different parameters under consideration. The study proposes a computation Offloading Decision Policy (ODP) framework to save battery lifetime, execution time, and memory utilization of IoT devices. This ODP framework estimates the execution time, energy consumption, and memory required for locally executing the task to be completed as well as when offloaded. The comparison between the loss function of locally and the remotely executed task performed. The proposed policy is compared with the traditional framework with no offloading at all and always full uploading. The results show improvement over traditional and other offloading frameworks. This technique applies to existing applications such as Smart Home, Industrial IoT, Intelligent traffic, Video Analytics, and Smart Healthcare delivers the power of AI. The ODP framework makes predictions for both the locally executed and offloaded versions of a task's execution time, energy use, and memory requirements. The outcomes demonstrate advancements above conventional and alternative offloading systems.

Keyword:

Computation offloading

Distributed architecture

Internet of thing

Edge computing

Resource management