You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am assuming smokesignal doesn't simply connect peers directly to each other, so a single peer will not be aware of every other peer on the network.
Given this assumption, how easy / difficult would it be to obtain a peer count for the whole network? Is this a tractable problem, or are my assumptions about the topology of a smokesignal network incorrect?
Thanks!
The text was updated successfully, but these errors were encountered:
Ah. No, only immediate neighbors are called peers here. Counting all
connected nodes is possible, though.
You could, for example, tell all nodes to connect to a specific port of
your node (e.g. where a simple tcp server is listening), so you can
count them.
You could also take a graph tracing approach: Send a message
COUNT(messageId) to all your peers. Nodes that already got the message
return a COUNTED(0). A node that has not got the message yet, forwards
it to all it's peers, adds the returned numbers and returns a
COUNTED(returnedTotal + 1).
If you're the node that started it, returnedTotal+1 is the total number
of nodes in the network.
Personally I'd like to avoid having a central server acting as a network overseer. Ideally I'd like it to be as decentralized as possible, so the graph tracing approach you mentioned might be the best way to go. Thanks for the information, hopefully I can put it to good use!
I am assuming smokesignal doesn't simply connect peers directly to each other, so a single peer will not be aware of every other peer on the network.
Given this assumption, how easy / difficult would it be to obtain a peer count for the whole network? Is this a tractable problem, or are my assumptions about the topology of a smokesignal network incorrect?
Thanks!
The text was updated successfully, but these errors were encountered: