Turbocharging AKS networking with Calico eBPF
A single Kubernetes cluster expends a small percentage of its total available assigned resources on delivering in-cluster networking. We don’t have to be satisfied with this, though—achieving the lowest possible overhead can provide significant cost savings and performance improvements if you are running network-intensive workloads. This article explores and explains the improvements that can be achieved in Microsoft Azure using Calico eBPF in AKS, including reducing CPU usage, decreasing complexity, enabling easier compliance and troubleshooting, and more.
Before going into details about how exactly Calico takes advantage of eBPF, it is important to note that in the context of this article, Calico is viewed as an additional networking layer on top of Azure CNI, providing functionality that turbocharges its performance. In particular, the standard instructions for installing Calico’s network policy engine with AKS use a version of Calico that pre-dates eBPF mode.
Accelerating network performance
Test methodology
To show how Calico accelerates AKS network performance using eBPF, the Calico team ran a series of network performance benchmarks based on the k8s-bench-suite. These performance benchmarks compared the latest Calico eBPF data plane (using the iptables data plane) with a vanilla AKS cluster.
Tests were run using Standard_D2s_v3 nodes, which are a Continue reading


Seamlessly, every single day, we wake up and check our health statistics in smart watches, scan QR codes to validate information, pay using credit cards in different locations, use surveillance cameras to record our neighborhoods, and connect our smartphones to distributed WiFi access points in our restaurants or coffee shops. 


