I’m developing an application on Isaac Sim. Since my laptop doesn’t have GPU, I intend to build the application on a cloud server. Isaac Sim’s documentation specifies that there are only some cloud providers (e.g. AWS, Azure, Tencent, …) for cloud deployment. That’s just for deployment, not development right? Is the meaning of deployment and development similar in this context? Would I be correct to assume that I can develop Isaac Sim application on any cloud provider given that they offer suitable services, like a minimum RTX-3070 with 32GB RAM and 8 VRAM or higher. Or somehow, Isaac Sim development and deployment are only possible on cloud servers of mentioned providers? On the cloud server, I should follow the Container Installation instructions then stream Isaac Sim to my laptop if the cloud service doesn’t include a GUI right?
The deployment should mean making Isaac Sim applications available on the Cloud. Yes, you can just follow the instructions from the specific cloud service documentation, which refers to the container page.
Thank you for your response! I understand that “deployment” refers to making Isaac Sim applications available on the cloud, and I can follow the Container Installation instructions for setup. However, I’m running into a specific issue with streaming Isaac Sim to my laptop, and I’d like some clarification on how this works with cloud providers—especially since my laptop and the cloud server aren’t on the same network.
I’m currently developing an Isaac Sim application (version 4.5.0) on a Vast.ai cloud server (RTX 4090, 24 GB VRAM, Ubuntu 22.04 Docker container) because my laptop (Ubuntu 22.04, no GPU) can’t run it locally. I’ve followed the Container Installation instructions, and I get the success message in the logs: [194.176s] Isaac Sim Full Streaming App is loaded after running ./runheadless.sh -v. However, when I try to stream it to the Isaac Sim WebRTC Streaming Client on my laptop using the server’s public IP (216.144.x.x:48338, mapped to port 3010), I get a blank screen. I suspect this is because my laptop and the Vast.ai server aren’t on the same network, as the documentation notes: “Isaac Sim WebRTC Streaming Client is recommended to be used within the same network as an Isaac Sim headless instance.”
So I have some questions about cloud deployment and streaming:
The documentation lists AWS, Azure, Tencent, etc., as supported cloud providers for deployment. If I deploy Isaac Sim on one of these (e.g., an AWS EC2 instance with an RTX GPU), the server still wouldn’t be on the same network as my laptop. How does streaming work in those cases? Does the WebRTC Streaming Client connect seamlessly to these providers over the internet, or is additional configuration (e.g., STUN/TURN servers) required?
Since Vast.ai isn’t explicitly listed, am I doing something wrong with my setup? The server meets the hardware requirements (RTX 4090, 24 GB VRAM, 64-core CPU), and Isaac Sim loads successfully, but streaming fails. Is there a difference in how Vast.ai handles networking or WebRTC compared to AWS/Azure that I need to account for?
For development, can I use any cloud provider like Vast.ai, assuming it meets the hardware specs, or are there specific features in AWS/Azure/Tencent/etc that make streaming work out of the box that Vast.ai might lack?
I’d really appreciate guidance on what I might be missing—whether it’s a configuration step for WebRTC streaming over the internet or a limitation with Vast.ai. My goal is to develop and test my application on the cloud server, streaming the output to my laptop for real-time interaction. Any insights or recommended steps would be incredibly helpful!