Migration Route Selection for a Microservice in a Fog Computing Network

Abstract

Purpose. The new generation of communication networks has provided the basis for the realization of new classes of services, one of which are URLLC services. These services include concepts such as Tactile Internet. The transmission medium for URLLC requires high bandwidth and reliability of communication channels as well as ultra-low latency in the network. To cover these requirements, various technologies, such as fog computing and microservice architecture, are introduced into the architecture of networks. The latter is more often recommended to be used together with support for microservice migration between devices. A related goal is to determine the efficient route to the destination server. Methods. Conducting research on a model network with the collection of real load statistics from the network nodes and their analysis to determine the most efficient migration route for microservice according to the proposed algorithm based on these data. Novelty. The novelty of the presented method consists in adapting the classical algorithm of route selection in a graph to solve the actual problem of microservice migration route selection. Results. The parameters that need to be considered for efficient routing during microservice migration and the level of network architecture that should control the migration were defined. The results have shown that the proposed method is preferred over the AODV protocol. Practical relevance. It is quantitatively shown that using the proposed method the CPU resources consumption on the network nodes during microservice migration can be reduced total by 77 %, the RAM resources usage by 14 %, the percentage of lost packets on the used communication channels can be reduced total by 7 %, which would provide higher network reliability and increase its performance.