Примечание
Для доступа к этой странице требуется авторизация. Вы можете попробовать войти или изменить каталоги.
Для доступа к этой странице требуется авторизация. Вы можете попробовать изменить каталоги.
Azure AI Video Indexer enabled by Arc is an Azure Arc extension enabled service that runs video and audio analysis, and generative AI on edge devices. The solution is designed to run on Azure Arc enabled Kubernetes and supports many video formats, and assumes that one Video Indexer resource is mapped to one extension. It can transcribe spoken content in more than 35 source languages and supports translation to English. For a full list of supported languages, see Supported languages per scenario.
This article walks you through the steps required to enable Video Indexer as an Arc extension in your current infrastructure.
Prerequisites
Important
To successfully deploy the Azure AI Video Indexer extension, it's mandatory that your Azure subscription ID is approved in advance. You must first sign up using the form at Application for gated services.
- Create an Azure subscription and assign permissions to a user so they can create Azure resources.
- Create an Azure AI Video Indexer account. For more information about creating an account, see the Create a Video Indexer account tutorial.
- Create an Arc enabled Kubernetes cluster.
To use the Video Indexer extension, you need to have an externally facing endpoint, which can be either a DNS name or an IP address. The endpoint should be set as a secure transfer protocol (https:\\) and is used as the extension API endpoint. It also gets used by the Video Indexer web portal to communicate with the extension. We recommend that you use an ingress control to manage the endpoint.
Note
If the endpoint isn't publicly accessible, you can perform actions on the extension from the web portal only from the local network.
Minimum hardware requirements for this quickstart
This quickstart is designed to allow you to see the extension in action. Smaller resource sizes are suggested for you to work with in a test environment. For this quickstart, the minimum hardware requirements are:
- CPU: 16 cores
- Memory: 16 GB
The CPU in the nodes should support AVX2. Most newer CPUs support the extension, but it might not be supported in some older virtualization environments.
To view minimum hardware requirements in a production environment, see the Minimum hardware requirements in the overview article.
Minimum software requirements
| Component | Минимальные требования |
|---|---|
| Операционная система | Ubuntu 22.04 LTS or any Linux Compatible OS |
| Kubernetes | 1,29 |
| Azure CLI | 2.64.0 |
Parameter definitions
Here's a table of the parameters used to configure the extension.
| Parameter | Description |
|---|---|
| release-namespace | The Kubernetes namespace that the extension is installed into |
| cluster-name | The Kubernetes Azure Arc instance name |
| resource-group | The Kubernetes Azure Arc resource group name |
| version | Video Indexer Extension version, leave empty for latest |
| videoIndexer.accountId | Video Indexer account ID |
| videoIndexer.endpointUri | URL containing a DNS name or IP address to be used as the extension external endpoint |
| ViAi.gpu.enabled | Enable GPU usage for summarization |
| ViAi.gpu.tolerations.key | Maps the nodes where summarization runs with GPU. Convention is to set to nvidia.com/gpu |
| ViAi.gpu.nodeSelector.workload | Identifies the node selected for summarization. Задайте значение summarization. |
Prepare for deployment
During the deployment, the script asks for environment specific values. Have these values ready so you can copy and paste them when the script asks for them.
| Question | Value | Details |
|---|---|---|
| What is the Video Indexer account ID during deployment? | GUID | Your Video Indexer account ID |
| What is the Azure subscription ID during deployment? | GUID | Идентификатор подписки Azure |
| What is the name of the Video Indexer resource group during deployment? | string | The resource group name of your Video Indexer account |
| What is the name of the Video Indexer account during deployment? | string | Your Video Indexer account name |
Deploy with the Azure portal
- In the Azure portal, navigate to your Azure Arc-connected cluster.
- From the menu, select Extensions>+ Add>Azure AI Video Indexer Arc Extension.
- Select Create. The Create an AI Video Indexer extension screen appears.
- Configure the extension in Instance details:
- Select the subscription and resource group for your extension.
- Select the region and connected k8 cluster.
- Enter a name for your extension.
- Select the Azure AI Video Indexer Account that the extension connects to.
- Enter the cluster endpoint, either an IP address or DNS name to use as the API endpoint.
- Provide the storage class you want to use for the extension that's supported by your Kubernetes distribution. For example, if you're using AKS, you could use
azurefile-cli. For more information about predefined storage classes supported by AKS, see Storage Classes in AKS. If you're using other Kubernetes distributions, see your Kubernetes distribution documentation for predefined storage classes supported or the way you can provide your own. - Select a generative AI model to apply AI capabilities such as textual summarization on VI enabled by Arc. For more information, see Generative AI with Azure AI Video Indexer (VI).
- Выберите Просмотр и создание, а затем нажмите кнопку Создать.
Manual deployment
Use the sample deployment script to manually deploy the extension. Before you get started, consider Storage class.
Storage class - Video Indexer extension requires that a storage volume must be available on the Kubernetes cluster. The storage class needs to support ReadWriteMany. It's important to note that the indexing process is IO intensive, so the IOPS (input/output operations per second) of the storage volume has a significant effect on the duration of the process.
Important
If you're using a language model, you must label a node or a node pool with workload:summarization. The label is a key-value pair, the key is workload, and the value is summarization. The machine labeled with this label must have at least 32 CPUs (for production), and we strongly recommend that they're Intel CPUs (as opposed to AMD).
Tip
Read the How to connect your cluster to Azure Arc article for a complete walkthrough of the process.
Optional configuration
The extension default settings are set to handle the common workloads, for specific cases, the following parameters can be used to configure the resource allocation:
| Parameter | Default | Description |
|---|---|---|
| videoIndexer.webapi.resources.requests.cpu | 0.5 | The request number of cores for the web API pod |
| videoIndexer.webapi.resources.requests.mem | 4Gi | The request memory capacity for the web API pod |
| videoIndexer.webapi.resources.limits.cpu | 1 | The limits number of cores for the web API pod |
| videoIndexer.webapi.resources.limits.mem | 6Gi | The limits memory capacity for the web API pod |
| storage.storageClass | "" | The storage class to be used |
| storage.useExternalPvc | false | Determines whether an external PVC is used. When true, the VideoIndexer PVC isn't installed. |
| scaling.ai.maxReplicaCount | 20 | Sets the AI workload max pod scale (not including summarization) |
Дальнейшие действия
Review the Azure AI Video Indexer enabled by Arc samples.