Heterogeneous cloud-edge computing-oriented universal task unloading method

Abstract

The invention discloses a universal task unloading method for heterogeneous cloud-edge computing, which comprises the following steps of: firstly, modeling a task into a directed graph GT; obtaining calculation cost and communication cost of each task in different environments; secondly, randomly initializing an unloading partition for a task node in the GT, and defining the unloading partition as an initial partition; calculating a gain obtained by moving the task from the current partition to the opposite partition for each task according to the initial partition; and performing iterative optimization on the initial partition according to the gain to obtain a final partition, wherein the final partition corresponds to a final task unloading strategy. According to the method, the efficient task unloading strategy oriented to the heterogeneous cloud-edge environment is obtained by utilizing a method for obtaining the maximum gain through iteration, and the total cost for unloading the whole task application in the heterogeneous cloud-edge environment can be obviously reduced.

Description

technical field [0001] The invention relates to a general and efficient task offloading algorithm oriented to a heterogeneous cloud-side environment, and belongs to the field of cloud-side computing. Background technique [0002] Edge computing has received a lot of attention due to its scalability and low communication costs, which compensate for the cost of transferring large amounts of data from local devices to remote clouds. With the rapid development of the Internet of Things, 5G networks, and smart devices (such as smartphones, wearable devices, and virtual reality devices), people are concerned about how to solve the limited resources on IoT devices and reduce cloud task data transmission due to bandwidth limitations. High latency makes urgent demands. Cloud-edge computing has the advantages of edge computing and cloud computing. Plus, it's one of the most effective ways to address IoT device resource constraints and just offload the entire application to the cloud...

AI Technical Summary

Problems solved by technology

[0005] In the prior art, there has been research on task offloading algorithms for cloud computing or edge computing, but due to the diverse heterogeneous resources and complex network structures between the edge and the cloud and other limiting factors, the task offloading problem in cloud-edge computing is more challenging than that in cloud computing and edge computing. Therefore, it is necessary to directly apply existing cloud computing and edge computing algorithms to cloud-edge computing. unrealistic

[0006] However, most of the existing cloud-edge offloading strategies ignore the heterogeneity and asymmetry of communication costs in cloud-edge computing. Therefore, the present invention proposes An efficient cloud-edge task offload method that is more suitable for the needs of real-world communication environments

Images