Kube-VIM is a Kubernetes controller for NFV resources. It turns ETSI MANO requests (Or-Vi and Vi-Vnfm) into Kubernetes Custom Resources and manages their lifecycle, statelessly. Every managed resource lives as a Kubernetes object in etcd (KRM). Because it runs in Kubernetes and reuses standard components (KubeVirt, CDI, Kube-OVN, Multus), your VNFs land on the same network and the same resource pool as your CNFs, with no parallel infrastructure. It speaks the ETSI reference points over gRPC and REST/OpenAPI, so any ETSI-compatible NFV orchestrator can drive it.
KRM-native
Every managed resource is a Kubernetes object in etcd. Kube-VIM reconciles NFV resources like a native Kubernetes controller, with no external database and no drift.
ETSI VIM, any NFVO
Or-Vi and Vi-Vnfm over gRPC and REST, with a generated OpenAPI client. Standards-conformant to IFA 005/006 and SOL 013. Validated with OSM via an OSM connector.
Line-rate data plane
SR-IOV with hardware offload for near-line-rate throughput. Overlay, underlay, and SR-IOV networks today; vDPA and vhost-user planned.
Converged NFV
Run VNFs alongside your CNFs on the same hardware and the same network fabric. Convergence is a property of being Kubernetes-native, not a bolt-on.
Hyperconverged and portable
Runs on any Kubernetes distribution. Proven on Talos Linux in a hyperconverged edge topology.
Easy to deploy
A single Helm bundle ships Kube-VIM together with the components it drives. GitOps-friendly and simple to stand up.
Built on trusted OSS
Stands on KubeVirt, CDI, Kube-OVN, and Multus, best-in-class open source with no vendor, NIC, or distribution lock-in.
Observable
Prometheus metrics out of the box, with kubevim_*_info labels that correlate
backend series to ETSI resource IDs.
By telco, for telco, and open source
Kube-VIM is open source under Apache 2.0, backed and supported by Retelith, a telecom software company building Kubernetes-native infrastructure for network operators.