A qoe-aware edge computing task scheduling method

Abstract

A QoE-aware edge computing task scheduling method. The steps are: the router periodically collects device resource status information and network status information. When the router receives a task scheduling request, it first predicts the execution time of the task on different devices according to the prediction model. On this basis, an edge computing task scheduling model is established based on device resource status information, network status information, and task data, and a linear programming model solver is used to solve the task scheduling model to obtain task scheduling results. Finally, based on the task scheduling result, the router is responsible for distributing the task to the corresponding edge device / cloud server for execution, and returning the result to the task requesting device.

Description

technical field [0001] The invention relates to a method for scheduling edge computing tasks based on QoE perception. Background technique [0002] The development of Internet of Things technology has brought massive heterogeneous data and a large number of Internet of Things devices. In the field of video surveillance, an 8-megapixel camera generates 3.6GB of data per hour, and the amount of data in a city reaches hundreds of petabytes a month. Currently, most IoT data processing relies on cloud computing technology. The increase in data volume has brought new challenges to cloud computing. First, the computing power of cloud computing cannot process massive amounts of data in real time. In the Internet of Things environment, the amount of data is growing explosively. Although cloud computing technology can improve the computing power of the system by increasing the number of devices in the cluster, it still cannot meet the requirements of real-time computing of massive ...

AI Technical Summary

Problems solved by technology

First, the computing power of cloud computing cannot process massive amounts of data in real time

In the Internet of Things environment, the amount of data is growing explosively. Although cloud computing technology can improve the computing power of the system by increasing the number of devices in the cluster, it still cannot meet the requirements of real-time computing of massive data.

Second, a large amount of data is transmitted from the network device to the cloud, which increases the network load and causes excessive network delay.

The traditional task scheduling algorithm does not consider the impact of dynamic network status changes and differences in device resources on task execution time. There are certain defects that may lead to unbalanced device loads and have a certain impact on task execution efficiency.

Images