How to Enable Inter-Instance Communication or Set Up a Custom Cluster on Lambda Labs?

post by johnnydaidai on Dec 3, 2024

Dear Lambda Labs Support Team,

I am currently working on a project that requires multiple instances to communicate seamlessly for distributed computing and clustering purposes. However, I have encountered some challenges and would like to clarify a few points regarding inter-instance communication on your platform.

Challenges:

• I have created multiple instances on Lambda Labs, but they appear to lack a direct communication interface (e.g., private networking or shared network interfaces).
• There doesn’t seem to be a built-in way to cluster these instances for low-latency communication or distributed workloads.

Questions:

1. Is there an option to enable private networking for instances within the same account or region?
2. Does Lambda Labs offer a built-in clustering feature for multiple instances, or should I configure a custom solution (e.g., VPN or Docker overlay network)?
3. Could you recommend best practices or provide documentation for setting up a cluster across Lambda Labs instances?

My use case involves distributed machine learning training, where low-latency communication is critical. If there is a recommended setup (e.g., enabling private IPs or configuring VPC-like networking), I would greatly appreciate your guidance.

Thank you for your time and support. Please let me know if additional information is required.

Best regards,