Artifact Manifest#
This section provides a comprehensive list of all components required for NVIDIA Cloud Functions (NVCF) Self-Hosted deployment for basic inference. Additional components are needed for Low Latency Streaming (Simulation).
Tip
Deploying to AWS with ECR? If you’re using the Terraform Installation to deploy to Amazon EKS, you can skip manual artifact mirroring entirely. Set create_sm_ecr_repos = true in your terraform.tfvars and Terraform will automatically mirror all required images and Helm charts to ECR. See Recommended for ECR Users: Automated ECR Mirroring for details.
Artifacts Overview#
The following tables list all artifacts required for an inference-only self-hosted NVCF deployment, organized by category, with their container images, Helm charts, and other resources.
Warning
Early Access (EA) Version Policy
During Early Access, artifact versions are updated frequently. The versions shown for Infrastructure Components are stable references, but all other components should use the latest published version from NGC.
To find the latest versions:
First, ensure you have the latest version of the NGC CLI installed and configured.
# List available versions for any container image
ngc registry image list "0833294136851237/nvcf-ncp-staging/<artifact-name>:*"
# For Helm charts (OCI-compliant charts are stored in the container registry)
ngc registry image list "0833294136851237/nvcf-ncp-staging/<chart-name>:*"
Note
OCI Helm Charts
All Helm charts listed in this manifest are OCI-compliant and stored in the NGC container registry (not the traditional Helm chart registry). This means:
Charts are pulled using
oci://URLs:helm pull oci://nvcr.io/0833294136851237/nvcf-ncp-staging/<chart-name> --version <version>Charts are listed using the image registry command:
ngc registry image listWhen mirroring to private registries (e.g., ECR), use container image tools like
skopeoorhelm push/pullwith OCI support
Important
Some supporting components such as the GPU Operator, OpenBao, NATS, Cassandra, etc. can alternatively be pulled directly from public NGC Catalog or other public opensource repositories if desired.
Artifact Registry Paths#
Infrastructure Components#
Core infrastructure services including NATS for messaging, Cassandra for data storage, and OpenBao for secret management.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
nats-box |
|
Image |
nats-server |
|
Image |
nats-server-config-reloader |
|
Chart (OCI) |
helm-nvcf-nats |
|
Image |
bitnami-cassandra |
|
Image |
nvcf-cassandra-migrations |
|
Chart (OCI) |
helm-nvcf-cassandra |
|
Image |
nvcf-openbao |
|
Image |
nvcf-openbao-migrations |
|
Chart (OCI) |
helm-nvcf-openbao-server |
|
Image |
oss-vault-k8s |
|
Control Plane Components#
Services that manage the NVCF platform including API gateway, deployment orchestration, invocation handling, and security services.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
spot |
|
Image |
strap |
|
Chart (OCI) |
helm-nvcf-api |
|
Chart (OCI) |
helm-nvcf-sis |
|
Image |
nvcf-grpc-proxy |
|
Chart (OCI) |
helm-nvcf-grpc-proxy |
|
Image |
nvcf-invocation-service |
|
Chart (OCI) |
helm-nvcf-invocation-service |
|
Image |
ess-api |
|
Chart (OCI) |
helm-nvcf-ess-api |
|
Image |
notary-service |
|
Chart (OCI) |
helm-nvcf-notary-service |
|
Image |
nv-api-keys |
|
Chart (OCI) |
helm-nvcf-api-keys |
|
GPU Workload Components#
Components that run on GPU nodes to manage function execution, including the NVCA operator and supporting containers.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
nvca |
|
Image |
nvca-operator |
|
Chart (OCI) |
helm-nvca-operator |
|
Image |
nvcf_worker_utils |
|
Image |
nvcf_worker_init |
|
Image |
nvcf_worker_niclls |
|
Image |
ess-agent |
|
Image |
nvcf-image-credential-helper |
|
Supporting Components#
Additional utilities and helper services required for the platform, including the NVIDIA GPU Operator for GPU node management.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
alpine-k8s |
|
Chart (HTTP) |
gpu-operator |
|
Image |
gpu-operator-validator |
|
Image |
k8s-device-plugin |
|
Chart (HTTP) |
ebs-csi-driver |
|
Chart (HTTP) |
csi-driver-smb |
|
Image |
load_tester_supreme |
|
Reference Architecture Components#
Optional components for the reference deployment architecture.
Type |
Component Name |
Full Path |
|---|---|---|
Chart (OCI) |
nvcf-gateway-routes |
|
Image |
admin-token-issuer-proxy |
|
Chart (OCI) |
helm-admin-token-issuer-proxy |
|
Observability Components#
Optional example components for monitoring and observability. These are provided as reference implementations only and are not intended for production use. See Example Dashboards Deployment for deployment instructions.
Type |
Component Name |
Full Path |
|---|---|---|
Chart (OCI) |
nvcf-observability-reference-stack |
|
Chart (OCI) |
nvcf-example-dashboards |
|
Container Caching Components#
Optional components for accelerating container image pulls across all workload types.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
nvcf-container-cache |
|
Chart (OCI) |
helm-nvcf-container-cache |
|
Image |
nvcf-proxy-tls-certs |
|
Image |
cuda |
|
Simulation Caching Components#
Optional caching components for Low Latency Streaming (LLS) and simulation workloads, including shader caching, derived data caching, and USD content caching.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
gxcache-webhook |
|
Image |
gxcache-init |
|
Image |
gxcache-service |
|
Chart (OCI) |
helm-gxcache |
|
Image |
ddcs-dist-kv |
|
Chart (OCI) |
helm-ddcs |
|
Image |
usd-content-cache |
|
Chart (OCI) |
helm-usd-content-cache |
|
Storage API Components#
Optional components for USD Storage API functionality used in simulation workloads.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
storage-service |
|
Chart (OCI) |
helm-storage-service |
|
Image |
simple-nginx |
|
Chart (OCI) |
helm-discovery-service |
|
Low Latency Streaming (LLS) Components#
Components for Low Latency Streaming functionality.
Type |
Component Name |
Full Path |
|---|---|---|
Image |
streaming-proxy |
|
Chart (OCI) |
gdn-streaming |
|
Deployment Resources#
Terraform and Helmfile resources for automated deployment.
Type |
Component Name |
Full Path |
|---|---|---|
Resource |
nvcf-base |
|
Resource |
nvcf-self-managed-stack |
|
Resource |
nvcf-cli |
|
Component Descriptions#
Infrastructure Components#
Component Name |
Description |
|---|---|
nats-box |
NATS utility container for debugging and administration |
nats-server |
Pub Sub Messages, used for Function Invocation and Deployment |
nats-server-config-reloader |
Configuration reloader for NATS server |
helm-nvcf-nats |
Helm chart for NATS deployment |
bitnami-cassandra |
Database for Account, Function and Cluster Management |
nvcf-cassandra-migrations |
Database migration scripts for Cassandra |
helm-nvcf-cassandra |
Helm chart for Cassandra deployment |
nvcf-openbao |
Secret management (OpenBao/Vault) |
nvcf-openbao-migrations |
Migration scripts for OpenBao |
helm-nvcf-openbao-server |
OpenBao Helm chart |
oss-vault-k8s |
Kubernetes integration for secret management |
Control Plane Components#
Component Name |
Description |
|---|---|
spot |
Spot Instance Service (SIS) - Manages deployments, cluster and instance state |
strap |
NVCF API service, refer to API for full API specification |
helm-nvcf-api |
Helm chart for NVCF API service |
helm-nvcf-sis |
Helm chart for Spot Instance Service |
nvcf-grpc-proxy |
Used for bi-directional communication and state management |
helm-nvcf-grpc-proxy |
Helm chart for GRPC Proxy deployment |
nvcf-invocation-service |
Handles stateless HTTP Function invocation requests |
helm-nvcf-invocation-service |
Helm chart for Invocation Service |
ess-api |
Encrypted Secrets Service - Used for application secret injection |
helm-nvcf-ess-api |
Helm chart for ESS API |
notary-service |
Used to sign and validate Functions and nodes |
helm-nvcf-notary-service |
Helm chart for Notary Service |
nv-api-keys |
API Key generation and management |
helm-nvcf-api-keys |
Helm chart for API Keys service |
GPU Workload Components#
Component Name |
Description |
|---|---|
nvca |
Performs the registration of the cluster and deployment orchestration in-cluster |
nvca-operator (image) |
Kubernetes operator for NVCA deployment and management |
nvca-operator (chart) |
Helm chart for NVCA operator deployment |
nvcf_worker_utils |
Acts as a proxy to NATS from the user’s application |
nvcf_worker_init |
Setup & Resource loading on deployment for the users application |
nvcf_worker_niclls |
NIC LLS worker component for low latency streaming workloads |
ess-agent |
Injects User Secrets |
nvcf-image-credential-helper |
Helper for managing container image credentials |
Supporting Components#
Component Name |
Description |
|---|---|
alpine-k8s |
Kubernetes utility container |
gpu-operator |
NVIDIA GPU Operator for dynamic GPU discovery - Can also pull directly from public NGC Catalog |
gpu-operator-validator |
GPU Operator validation component |
k8s-device-plugin |
Kubernetes device plugin for GPU support |
ebs-csi-driver |
AWS EBS CSI Driver for persistent volume provisioning on EKS |
csi-driver-smb |
CSI Driver for SMB/CIFS file shares |
Reference Architecture Components#
Component Name |
Description |
|---|---|
nvcf-gateway-routes |
Gateway routing configuration for reference architecture |
admin-token-issuer-proxy |
Admin token management proxy |
helm-admin-token-issuer-proxy |
Helm chart for admin token issuer proxy |
Observability Components#
Component Name |
Description |
|---|---|
nvcf-observability-reference-stack |
Reference observability backend (Prometheus, Grafana, Loki, Tempo, OpenTelemetry Collector) |
nvcf-example-dashboards |
Pre-configured Grafana dashboards for NVCF control-plane metrics |
Container Caching Components#
Component Name |
Description |
|---|---|
nvcf-container-cache |
Accelerates container image pulls by caching layers locally on nodes |
helm-nvcf-container-cache |
Helm chart for container cache deployment |
nvcf-proxy-tls-certs |
TLS certificate management for container cache proxy |
cuda |
CUDA base image used as a dependency for container cache |
Simulation Caching Components#
Component Name |
Description |
|---|---|
gxcache-webhook |
Shader cache webhook for intercepting and caching shader compilation requests |
gxcache-init |
Init container for shader cache setup |
gxcache-service |
Backend service for shader cache storage and retrieval |
helm-gxcache |
Helm chart for deploying the complete shader cache stack |
ddcs-dist-kv |
Derived Data Cache Service - caches computed/derived data for simulation workloads |
helm-ddcs |
Helm chart for DDCS deployment |
usd-content-cache |
USD Content Cache - caches Universal Scene Description assets for streaming |
helm-usd-content-cache |
Helm chart for USD Content Cache deployment |
Storage API Components#
Component Name |
Description |
|---|---|
storage-service |
USD Storage Service for managing assets in simulation workloads |
helm-storage-service |
Helm chart for Storage Service deployment |
simple-nginx |
Simple NGINX container for Storage API routing |
helm-discovery-service |
Helm chart for Storage API Discovery Service |
Low Latency Streaming (LLS) Components#
Component Name |
Description |
|---|---|
streaming-proxy |
LLS Streaming Proxy Container |
gdn-streaming |
LLS Self-Hosted Helm Chart |
Deployment Resources#
Component Name |
Description |
|---|---|
nvcf-base |
Terraform configuration for NVCF quickstart deployment |
nvcf-self-managed-stack |
Helmfile bundle for self-managed stack deployment |
nvcf-cli |
Command-line interface for managing functions and deployments |