The video discusses the implementation of IPv6 eBPF clusters with Calico, providing an early preview of the feature. The speaker, Toas, explains the methodology of setting up an IPv6 cluster using eBPF in Calico, highlighting the limitations, such as the requirement for IPv6-only clusters and lack of dual-stack support in the preview version. Here is the methodology presented in a detailed bullet point list:
- Setting up multipass for the demo environment
- Editing network configurations to enable IPv6
- Running the control plane node and worker node
- Transferring the Kubernetes config file to the host computer
- Installing the Tigera operator with Calico using the evf data plane
- Creating Global Network Policies and allowing specific traffic within the cluster
- Checking the status of the eBPF data plane using various commands like `network`, `contract`, `routes`, and `policy`
### Methodology
- Setting up multipass for the demo environment
- Editing network configurations to enable IPv6
- Running the control plane node and worker node
- Transferring the Kubernetes config file to the host computer
- Installing the Tigera operator with Calico using the evf data plane
- Creating Global Network Policies and allowing specific traffic within the cluster
- Checking the status of the eBPF data plane using various commands like `network`, `contract`, `routes`, and `policy`
### Speakers
- Resa
- Toas
Notable Quotes
—
29:08
— « Let's start by running some codes. Uh, by the way, if somebody is curious to know how to run this, they can find everything here.
»
—
33:42
— « Let's secure it with um with a bunch of policies and then ask Tamas about ebpf policies and how they differ from IP table ones.
»
—
35:22
— « So at this point, Toas, I've enabled the workload and I've got the um policies inside my cluster but I want to verify if this cluster is actually running ebpf, right?
»
—
37:52
— « We're processing IPv6 packets and that we actually do process packets running in the ebpf mode. »