Getting Started

In this section we will deploy infrastructure to be able to deploy NAI.

We will deploy a jumphost to work with. We will deploy all the necessary tools on this jumphost. Once this is done, we will use this jumphost to deploy the NAI platform.

This guide covers two different scenarios for deploying Nutanix Enterprise AI [NAI] (previously known as GPT-In-A-Box).

1. The first scenario is a walkthrough on How-To Deploy Nutanix Enterprise AI (NAI) with Nutanix Kubernetes Platform (NKP), and
2. The second scenario covers the (soon to be deprecated) option on How-To Deploy GPT-In-A-Box v1 Nutanix Validated Design (NVD) with NKE .

Deprecation Notice!

GPT-In-A-Box v1 and NKE will be deprecated soon. We will be removing these options from the this site as soon as it is deprecated.

Each scenario goes through four phases to prepare the infrastructure on which you can deploy Nutanix Enterprise AI applications.

1. Prepare the workstation.
2. Deploy the jumphost
   1. Option A - Deploy NKP to deploy NAI
   2. Option B - Deploy NKE to deploy GPT-In-A-Box v1

Here is a diagram that shows the four phases.

stateDiagram-v2
  direction LR

  state PrepWorkstation {
    [*] --> GenrateRSAKeys
    GenrateRSAKeys --> InstallTofu
    InstallTofu --> InstallVSCode
    InstallVSCode --> [*]
  }
  state DeployJumpHost {
    [*] --> CreateCloudInit
    CreateCloudInit --> CreateJumpHostVM
    CreateJumpHostVM --> DeployNaiUtils
    DeployNaiUtils --> [*]
  }
  PrepWorkstation --> DeployJumpHost
  DeployJumpHost --> DeployNkp: Option A
  DeployNkp --> DeployNai
  DeployJumpHost --> DeployNke: Option B
  DeployNke --> DeployGiabGitOps

Deploy NAI with NKP

1. Prepare Local Development Workstation (Mac/Windows)
2. Deploy Jumphost VM
3. Deploy Nutanix Kubernetes Platform (NKP) Management Cluster
4. Deploy Nutanix Enterprise AI (NAI)

Here is a diagram that shows the workflow:

stateDiagram-v2
    direction LR

    state DeployNKP {
        [*] --> CreateNkpMachineImage
        CreateNkpMachineImage --> CreateNkpSelfManagedCluster
        CreateNkpSelfManagedCluster --> DeployGPUNodePool
        DeployGPUNodePool --> [*]
    }
    state DeployNai {
        [*] --> PrepareNai
        PrepareNai --> DeployNaiHelm 
        DeployNaiHelm --> [*]
    }

    [*] --> PrepWorkstation
    PrepWorkstation --> DeployJumpHost
    DeployJumpHost --> DeployNKP
    DeployNKP --> DeployNai
    DeployNai --> [*]

Deploy GPT-In-A-Box (v1) with NKE

1. Prepare your Local Development Workstation (Mac/Windows)
2. Deploy Jumphost VM
3. Deploy Nutanix Kubernetes Engine (NKE) - Management Cluster
4. Deploy Nutanix Kubernetes Engine (NKE) - Development Workload Cluster
5. Deploy Nutanix GPT-In-A-Box (v1) Validated Design Reference RAG Applications using Flux GitOps

Here is a diagram that shows the workflow:

stateDiagram-v2
    direction LR

    state DeployNKE {
        [*] --> CreateTofuWorkspaces
        CreateTofuWorkspaces --> CreateMgtK8SCluster
        CreateMgtK8SCluster --> CreateDevK8SCluster
        CreateDevK8SCluster --> DeployGPUNodePool
        DeployGPUNodePool --> [*]
    }
    state DeployGiabGitOps {
        [*] --> BootstrapFlux
        BootstrapFlux --> DeployAIApps
        DeployAIApps --> [*]
    }

    [*] --> PrepWorkstation
    PrepWorkstation --> DeployJumpHost
    DeployJumpHost --> DeployNKE
    DeployNKE --> DeployGiabGitOps
    DeployGiabGitOps --> [*]