I am trying to understand how the light client reduces dependence on RPC nodes. The typical RPC node uses the --ws-max-connections flag to set a specific limit on the number of web socket connections to that node. If you are using that RPC node as a light client bootnode, light client connections also come in on web socket. So wouldn't this limit on ws connections also limit the number of light clients booting off of that node? Is there a way to make that unlimited? Or will NGINX limit that itself regardless? Is the answer to scaling light client users simply to have lots of light client boot nodes? Which means having lots of RPC nodes?

1 Answer 1


Bootnodes don't necessarily need to be rpc nodes. Bootnodes are simply there to have a node join the network.

In addition, light clients will not only establish connections with RPC nodes. That's the cool thing regarding light clients. They join the network like any other node.

  • Neat! I understand that bootnodes for light clients must look something like this: using WS (example from Rococo) "/ip4/" ... in the case of my own boot nodes, would I use whatever the WS port is in place of the 30334 seen here? or would I still use the p2p port in that position? Thanks for your help! Jun 16, 2023 at 21:05
  • Last question... do those nodes need to be started in any special way such as with --listen-addr /ip4/ or something like that? Jun 16, 2023 at 21:27
  • I think you can use whatever port you want and check out this question Jun 20, 2023 at 9:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.