Dynamic Fog Computing Towards Green ICT

Abstract

Relevance: In the context of the growing fleet of data center equipment, the development of IMT-2020 networks and the imminent emergence of Telepresence services of IMT-2030 networks, a particularly relevant area of modern research is the search for non-trivial, non-standard approaches and solutions in the field of provision of computing and network resources. This article covers current issues in the infrastructure direction of IMT-2030 networks - dynamic fog computing. The contribution of this technology to improve the efficiency of used resources is considered, and current scenarios for IMT-2030 networks are presented. In particular, we study the problem of searching for a group of devices in the computing fog for subsequent migration of typical FaaS platform containers. Problem statement: Research on the joint use of serverless architecture and dynamic fog computing for efficient load distribution of telepresence services. Goal of the work: Research and development of an effective method for distributing a group of microservices in dynamic fog computing. Methods: the algorithms under study belong to the type of metaheuristic algorithms for solving multicriteria optimization problems. To test the method, a laboratory network segment was developed, which served as a generator of real data on the operation of the tested platforms under conditions of increasing load. Based on a series of experiments, data was collected that formed the basis for subsequent modeling of the proposed method, which in turn was implemented in the Python programming language. Result: Analysis of the results showed the effectiveness of the proposed method within the framework of the task, which ultimately makes it possible to make a decision on migration many times faster. Novelty: A model and method for serverless architecture have been developed for migrating groups of microservices to groups of fog computing devices, under conditions of their mobility, and a meta-heuristic algorithm of a pack of gray wolves has been used to determine a group of devices for subsequent migration of typical microservices. Practical significance: The developed model and method can be used in the implementation of fog Computing, in conditions of device mobility, including in order to achieve the requirements of promising Telepresence services.