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
Currently; each DSQ message that we receive is simply repropogated to all connected nodes. This is not ideal for a bandwidth usage perspective. Additionally, if a node explicitly doesn't care about dsq messages; there's no way to avoid sending it to them. Not only does this significantly increase receiving bandwidth, but also sending bandwidth. Instead we should use the inventory system!
A DS is ~142 bytes. If we have 10 peers; and we send and receive from each of them, then we are sending and receiving ~1.42kB / DSQ session. There are quite a lot of DSQ sessions that happen during a day.
Using inv; we'd send and receive an inv from each peer; so 37 * 10 = 370 bytes + (send and receive for 1 node) 142; this means in the new system we'd be sending receiving about 512 bytes per DSQ session. Additionally, this difference may be significantly higher for master nodes who may have upwards of 100 connections
The text was updated successfully, but these errors were encountered:
Currently; each DSQ message that we receive is simply repropogated to all connected nodes. This is not ideal for a bandwidth usage perspective. Additionally, if a node explicitly doesn't care about dsq messages; there's no way to avoid sending it to them. Not only does this significantly increase receiving bandwidth, but also sending bandwidth. Instead we should use the inventory system!
Calculations
A DS is ~142 bytes. If we have 10 peers; and we send and receive from each of them, then we are sending and receiving ~1.42kB / DSQ session. There are quite a lot of DSQ sessions that happen during a day.
Using inv; we'd send and receive an inv from each peer; so 37 * 10 = 370 bytes + (send and receive for 1 node) 142; this means in the new system we'd be sending receiving about 512 bytes per DSQ session. Additionally, this difference may be significantly higher for master nodes who may have upwards of 100 connections
The text was updated successfully, but these errors were encountered: