Development of an automation algorithm for restoring the operational capability of information and communication networks
TL;DR
The aim of this study was to develop and evaluate the effectiveness of an algorithm for the automated recovery of network nodes in information and communication networks using virtualisation and automation tools and found that the proposed automated algorithm reduces the network recovery time by 8,5 times compared to the manual method.
Abstract
The aim of this study was to develop and evaluate the effectiveness of an algorithm for the automated recovery of network nodes in information and communication networks using virtualisation and automation tools. An analysis of existing approaches to network equipment redundancy, in particular the Virtual Router Redundancy Protocol and Hot Standby Router Protocol, was conducted, and their key limitations were identified: the inability to restore the router to its full functional state, limited scalability, and reliance on the human factor. It was determined that the average time for manual restoration of network nodes ranges from 10 minutes to 1 hour or more, depending on many factors and conditions. A comparative analysis of modern configuration management tools, including Ansible, Puppet, Chef, and SaltStack, was conducted. The analysis results justified the choice of Ansible, a specialised automation tool, due to its agentless architecture, use of the secure Secure Shell protocol, and “PUSH” data transfer model, which provided instant responses without prior configuration of target nodes. An algorithm for automated deployment of a virtual router analogue in the Proxmox VE hypervisor environment, when the special Zabbix monitoring software detects a physical equipment failure, has been developed. Centralised management of the recovery process has been implemented via the Jenkins automation server, which receives WebHook signals from Zabbix and initiates the execution of Ansible scripts. The algorithm included sequential stages: failure detection, creation of a new virtual machine, application of the current configuration from the backup storage, and verification of the health of the restored node. To quantify effectiveness, an additive time model and a recovery-acceleration intensity coefficient were applied. It was found that the proposed automated algorithm reduces the network recovery time by 8,5 times compared to the manual method from 30-60 minutes to 3-4 minutes. The proposed algorithm was applicable as a supplementary redundancy layer