How do KubeEdge's edge nodes find nearby relay node?

Hi everyone, I am very interested in the distributed dynamic relay connectivity provided by the EdgeMesh subproject under KubeEdge.

I have been researching the EdgeMesh project and found in the documentation(here: https://github.com/kubeedge/edgemesh/blob/main/docs/proposal/edgemesh-ha-design-proposal.md) that it is possible to dynamically select nearby relay nodes based on the geographic distribution of cluster nodes to provide network connectivity services, and this implementation is achieved through DHT.

However, I believe that in EdgeMesh, the closest nodes found through DHT are defined by the XOR distance between node IDs, which might not reflect the actual physical distances between nodes. Perhaps there is something I'm misunderstanding. Could anyone clarify this for me?

Start with docs/proposal/edgemesh-ha-design-proposal.md and the DHT behavior described in the issue. Compare how the proposal defines a nearby relay with how XOR distance between node IDs is used; done means clarifying whether XOR distance represents physical proximity or explaining how geographic proximity is handled.

Written by the indexing model from the issue text.