# KubeCon NA NRI Talk Abstract

## Plugging Extra Functionality into Runtimes Using NRI

The Node Resource Interface is a common extension mechanism for OCI-compatible runtimes. It allows one to register plugins to extend the functionality of runtimes. NRI plugins can tap into the lifecycle events of pods and containers and track their state. Plugins can also adjust the configuration of containers, both in response to lifecycle events and otherwise. This allows one to implement custom logic for container management using NRI. Since NRI support is present in both CRI-O and containerd, custom logic implemented as an NRI plugin is usable with both of runtimes.

In this talk we take a brief look at how NRI works and what kind of data and controls it exposes to plugins. Then we discuss what kind of extra logic NRI allows one to implement using some existing plugins as examples.

## Target Track ???

With the new tracks, I'm not sure which one would be the most suitable.
Maybe Platform Engineering, or Operations + Performance...

# alternative idea
## NRI: One Plugin System to Rule Them All
Throughout the lifecycle of Kubernetes and the OCI, there have been several attempts to write a generic plugin system for container runtimes: oci-hooks, CDI, CRI-O's runtime handler hooks, and more. This causes skew in the ecosystem and an unclear direction for plugin authors.

With the advent of the Node Resource Interface (NRI), the ecosystem has a chance to rally behind one solution. As the NRI stabalizes and broad support has been added, it can begin to fulfill the many use cases previously fulfilled by the other solutions.

Join Krisztien and Peter as they walk through the capabilities of NRI, some new places NRI can be used, and ponder the future of a reunited plugin system for Kubernetes container runtimes.