Scalable Agentless Cloud Network Monitoring

摘要

With the growth of Infrastructure as a Service (IaaS) in cloud computing, and the Internet of Things (IoT), network monitoring of the global infrastructure will continue to expand, as cybersecurity will become of primary importance. Current network monitoring approaches, particularly those that are agent-based, are proprietary, costly, and may not scale. This paper proposes a lightweight, scalable agentless system which can be configured, setup, and begin monitoring network health within minutes. The architecture utilizes a polling script to query Simple Network Monitoring Protocol (SNMP) demon for metrics, Prometheus Time Series Database instances for storage, and a Grafana Dashboard for metric presentation and alerting. The system design removes the learning curve necessary for most open source monitoring systems, while still providing similar metrics. The approach demonstrated here shows the ease and security with which specific site network monitoring can be deployed in a cloud configuration. Data from the network monitoring can be forwarded to a central analysis site for assessment and correlation. The proposed system is scalable, extensible for use in most cloud hosting environments, and provides network managers with direct control over the monitoring process and dashboard.