C3 AI Air-Gapped Deployment Overview

C3 Agentic AI Platform supports deployment with restricted network egress. This configuration is sometimes called "air-gapped" but frequently does not mean full isolation of the network of the C3 AI cluster.

To manage the C3 Agentic AI Platform effectively so it functions efficiently, the C3 AI cluster requires network connectivity to central services like the following:

• Repositories for containers and code, whether public (Internet-based), customer-specific local repositories, or C3 AI repositories.

• Identity Provider (IdP), whether customer-managed or C3 AI-managed IdP.

• Data sources, whether customer-managed or public (Internet-based), from which Prod/Test environments get Prod/Test datasets.

• Infrastructure services, such as: storage, network, and ingress services.

• Supporting systems, whether C3 AI-specific or customer-specific bastion and continuous integration and continuous deployment (CICD) systems (such as, source controls, Jenkins, technology test cluster).

Special configurations are mostly needed for repositories. IdP integration is same for C3 AI no matter where the IdP is setup. This is also true for data sources.

If you need to create on-premises installation of the C3 Agentic AI Platform, contact your C3 AI Center of Excellence (CoE). Review supporting systems with your C3 AI CoE.

Copy link to this section

Overview of repository considerations for air-gapped deployments

For air-gapped deployments, additional configurations are necessary for repositories to facilitate management and functionality of the C3 Agentic AI Platform. C3 AI supports the following code repositories for air-gapped environments.

Copy link to this section

Python code repositories

C3 Agentic AI Platform uses Python code throughout the platform and Jupyter, supporting data science and machine learning (ML) use cases. Python packages are managed using Pip and Conda package managers and can be sourced from Internet locations, such as: repo.anaconda.com, conda.anaconda.org, repo.continuum.io, and pypi.org.

If you cannot access those repositories, you will need to do the following:

1. Create your own local repository cache with tools like Sonatype Nexus.

   See Example Configuration to Use Sonatype Nexus as a Repository Server.

2. Configure C3 AI Cluster to point to this local repository.

   See Configure C3 Agentic AI Platform to Use Local Conda and PyPI Repositories in an Air-Gapped Environment.

Copy link to this section

Node.js code repositories

C3 AI applications use Node.js code for UI. The npm package manager provides Node.js packages and can source the packages from Internet locations, such as: registry.npmjs.org, nodejs.org, and npmjs.com.

If you cannot access those repositories, you will need to do the following:

1. Create your own local repository cache with tools like Sonatype Nexus.

   See Example Configuration to Use Sonatype Nexus as a Repository Server.

2. Configure Microsoft Visual Studio (VS) Code to point to this local repository.

   See Setup Application Development in an Air-Gapped Environment.

3. Configure the C3 AI Cluster to point to this local repository.

   See Configure the C3 Agentic AI Platform to Bundle in an Air-Gapped Environment.

Copy link to this section

C3 AI repositories for container images

C3 AI Cluster installation is comprised of HashiCorp Terraform modules to create the infrastructure, Helm charts to define the C3 AI Ops Operators, and Helm charts to define the runtime cluster of C3 AI servers, Jupyter, Cassandra, and Postgres. These HashiCorp Terraform modules and Helm charts are sourced from Internet-based C3 AI repositories, such as: jfrog.c3.ai for the HashiCorp Terraform modules, and registry.c3.ai for the Helm charts and container images.

Access of C3 AI repositories for container images, Helm charts, and HashiCorp Terraform modules is dependent on whether the deployment is a C3 AI-managed or customer-managed deployment. Contact the C3 AI CoE for more information.

For air-gapped environments, since access to the Internet-based repositories is restricted, you will need to do the following:

1. Create your own local repository cache with tools like Sonatype Nexus. See Configure Sonatype Nexus as a Repository Server for an example.

2. Download the HashiCorp Terraform modules for C3 AI infrastructure creation and C3 Agentic AI Platform deployment in a zip file and deploy from them. See the C3 AI ClusterBOM to validate the versions and dependencies.

3. Download the Helm charts for C3 AI infrastructure creation and C3 Agentic AI Platform deployment in a zip file and deploy from them. See the C3 AI ClusterBOM to validate the versions and dependencies.

4. Configure the C3 Agentic AI Platform installation to point to the local repository. See Use C3 AI Containers and Images in an Air-Gapped Environment for more information.

See also Overview of C3 AI Supported Deployments.

Copy link to this section

See also

For more information on release and cloud specific installation requirements, see the Supported Platform Versions page.