Measuring control plane performance
Most network engineers are familiar with tools such as ping, traceroute, and iperf to verify network data plane connectivity, reachability, and throughput. At the same time, control plane performance often remains a black box, and we can only assume how long it takes for our network to re-converge. In this section, we aim to address this problem by building a control plane telemetry solution.
Modern control plane protocols, such as BGP, distribute large volumes of information from IP routes to MAC addresses and flow definitions. As the size of our networks grows, so does the churn rate of the control plane state, with users, VMs, and applications constantly moving between different locations and network segments. Hence, it’s critical to have visibility of how well our control plane performs to troubleshoot network issues and take any preemptive actions.
The next code example covers the telemetry processing pipeline we built to monitor...