Replies: 1 comment
-
I think the main limiting factor is the rate limit applied to DHT nodes. This defaults to 8 kB/s in libtorrent. i.e. the total number of requests that can be handled (and responded to) is proportional to the total number of nodes on the network (assuming node-IDs and keys are uniformly distributed). |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
After searching google for some time I have found there have been a lot of conversations with respect to "DHT Pressure" when performing excessive activity on the DHT. This has been mostly around
get_message
mutable/immutable data, but it has been brought up withfind_node
also. What I cannot seem to find is a quantification of what the boundaries are. I assume that DHT operations will take some level of network IO that will grow with traversal. There would also be CPU cycles on the peers to process those operations. Is there a way to determine the impact of 100 peers performing mutable data get/put operations on each other? How does that change if each peer has a mutable entry, and every other peer will query the other single mutable entry? Likewise, how does that change if only half the peers are looking for another single mutable entry? Does this grow linearly with the number of peers, or is it exponential/geometric/etc?TL;DR; How can I figure out the limit of DHT operations before performance deteriorates for peers?
Beta Was this translation helpful? Give feedback.
All reactions