Self-adaptive container migration method under fog computing environment

Abstract

The invention provides a self-adaptive container migration method under a fog computing environment. The method comprises the following steps that a container-based fog computing framework is established, wherein the container is located on a fog node, a mobile application is located on the body of a user body, and a task of the user is executed in the container; modeling is conducted on a targetmigration of the container under a fog computing scene, wherein the migration target comprises time delay, power consumption and migration cost; a state space and an action space are set, a reward function is defined, and a Q iterated function is set; dimensionality reduction is carried out on the state space through a deep neural network; dimensionality reduction of the action space is achieved through optimization of action selection; finally, a prototype of a self-adaptive migration system of the container is achieved, and the whole process is verified. By means of the provided self-adaptive container migration method under the fog computing environment, resources in fog computing can be better planned, the time delay between the user and the fog node is reduced, and the energy consumption expenditure of the fog node is reduced.

Description

technical field [0001] The invention belongs to the field of fog computing in computer networks, and relates to methods such as fog computing, mobile edge computing, reinforcement learning, and deep reinforcement learning, and in particular to an adaptive container migration method in a fog computing environment. Background technique [0002] Fog computing is emerging as a promising computing paradigm in recent years, providing a flexible architecture to support distributed region-specific, domain-specific applications with cloud computing-like quality of service. Fog computing deploys massive lightweight computing and storage infrastructures (called fog nodes) near mobile users. In this way, the mobile application can be distributed to the appropriate fog node to shorten the user's access delay to the application. In addition, fog nodes are flexible and scalable, and can support the mobility of mobile users. [0003] There are few existing techniques for container migrati...

AI Technical Summary

Problems solved by technology

[0005] Directly migrating the method of virtual machine management in the data center to the fog computing environment will bring a series of problems, including the disaster of dimensionality caused by high-dimensional state space and action space, and mobile users are not considered in the modeling process. Mobility problems, resulting in delay problems in mobile scenarios cannot be well resolved

[0006] However, the existing task scheduling methods for fog computing only consider the situation of a single user when establishing the state space, and do not consider the actual multi-user situation.

And it is assumed that the transition probability between states is fixed, but the transition probability between states is unknown in the actual situation

Images