Deployment of visual computing applications based on cloud XR platform

At 14:00 on July 20, 2022, the new product launch of the cloud XR platform launched by Alibaba Cloud in conjunction with parallel cloud was officially launched. Alibaba Cloud elastic computing product solution architect Jin Yunlong brought a share entitled "Visual computing application deployment based on the cloud XR platform".

Alibaba Cloud XR platform provides computing power scheduling, 3D real-time rendering, audio and video coding and streaming, application publishing and user management and other services. Developers can host application publishing to the cloud XR platform, and users can use multiple terminals such as AR/VR/MR to join the cloud XR platform and interact with and experience 3D applications. The following is the content of his speech for everyone to read:

Application delivery and deployment of computing nest mode

The cloud XR platform is deployed in the way of computing nest access. The computing nest service is a PaaS platform that Alibaba Cloud opens to enterprise application service providers and other customers.

Service providers can publish private deployment services on the computing nest and provide their customers with the ability to deploy software on the cloud with one click. At the same time, it also supports the full hosting mode service, enabling service providers to host customer resources. The XR platform for this demonstration has completed the docking with the computing nest.

Compared with the traditional delivery mode, the computing nest can deliver and deploy applications more easily and quickly, greatly reducing the delivery difficulty and shortening the delivery and deployment cycle. The traditional interaction mode usually submits the configuration list of resources, then manually creates resources according to the list, provides the application deployment manual, and finally installs and deploys. After docking with the computing nest, a service-oriented subscription mode can be created, which can greatly improve the overall efficiency of interactive deployment. Through the computing nest, cloud XR visual computing application deployment can be quickly built and completed in 5 minutes.

Cloud XR visual computing application deployment process

The specific steps are as follows:

① Give basic permissions to the console interface of the computing nest, including ECS permissions, VPC permissions, ROS permissions, etc.

② Obtain the permissions of the computing nest, and obtain the deployment links and service permissions of the computing nest.

③ To deploy nodes and configurations, you must first select the target region for deployment, and then select the specifications, disk size, bandwidth, etc. of rendering nodes and management nodes. You also need to set some basic configurations.

④ Create a service and wait for the deployment to complete. The deployment of the application can be completed in about 2-3 minutes. The underlying resource payment mode is very flexible, which can provide pay-as-you-go, yearly, monthly and other methods.

⑤ Add an application on the XR platform, enter the management interface of the deployed application, upload and save the locally prepared 3D application deployment.

⑥ Application opening and experience. Log in to the application display page, enter the application and experience, and exit after the experience. Resources can be flexibly processed, reserved or released.

Demo demonstration: deploy cloud XR real-time rendering platform through computing nest

Enter the console interface of the computing nest. The control menu is on the left. In "My Services", you can see that the third-party services of the cloud XR real-time rendering platform have been connected with the computing nest. The steps required to deploy the cloud XR real-time rendering platform are listed on the right.

First, select the region, set the number of nodes to 1, and select the lightweight GPU instance vgn6i-vws as the instance specification.

The system disk type is essd cloud disk. The system disk size can be flexibly configured from 40G-500G. Select No to create a data disk. The management node uses the CPU instance and selects the shared standard s6 with 8G memory. The system disk type is essd cloud disk, the size is set to 40G, and no data disk is created.

Configure a VPC for both render nodes and management nodes. If there is no VPC, click "New VPC" to create a VPC. Then select the recommended zone, select the switch and configure the password. The payment mode is "Pay as you go", the traffic payment is "Pay as you go", and the public network bandwidth is set to 50G. After completing the above settings, click "Create", confirm that the configuration is correct, and then click "Next" to make payment to enter the deployment process.

Deployment takes 2-3 minutes. You can click the instance ID in the service instance management to view the details of the creation process.

The overview includes events, service resource information, basic information and other configurations.

On the resource side, you can see that the management node and rendering node you just created are starting, configuring, and deploying.

On the monitoring interface side, the XR real-time rendering platform will display monitoring information after deployment. On the deployment log side, you can view the logs during deployment.

After deployment, click the service instance ID to view its resource creation and show that the resource is already running.

Log in to the real-time rendering platform, and the GPU and other monitoring data are displayed under the resource monitoring of the data center, which has been published by the rendering node.

Next, publish the application. Click "Add Application" in "Application Management" - "Application List" to upload the prepared application.

Name the application and select the execution path to publish the XR application in the real-time rendering platform.

Find the published application in the release list, and run it to experience the application. That's what I share today.

Extension

In the process of landing the cloud XR platform, Alibaba Cloud mainly provides super cloud computing power and self-developed streaming media coding technology. At the platform level, parallel cloud efficiently encapsulates rendering, coding and streaming services, integrates Alibaba Cloud's underlying IaaS capabilities, and supports the NVIDIA CloudXR platform.