1 2015-01-24 00:16:00 <Luke-Jr> hm, the p2p protocol reports peer count. I wonder if we should prefer fewer-peer nodes or something, or if that just invites lying
 2 2015-01-24 00:20:22 <phantomcircuit> Luke-Jr, does it really?
 3 2015-01-24 00:20:38 <Luke-Jr> I suppose maybe it's local peers
 4 2015-01-24 00:20:41 <Luke-Jr> 2015-01-23 07:12:18 receive version message: /Satoshi:0.10.0/ljr:20150112/: version 70002, blocks=134441, us=0.0.0.0:0, peer=16
 5 2015-01-24 00:20:48 <Luke-Jr> oh, it's the peer id I bet
 6 2015-01-24 00:20:54 <phantomcircuit> yeah that's the peer id
 7 2015-01-24 00:21:14 <phantomcircuit> iirc you cant only guess based on the "addr" message
 8 2015-01-24 00:38:23 <sipa> warren: that's not the point
 9 2015-01-24 00:38:54 <sipa> you need as many blocks queued as can be in flight
10 2015-01-24 00:39:51 <sipa> if your latency is 1s and it takes 0.1s to send/receive/process a block, you need a queue of 10 to keep the link optimally full
11 2015-01-24 00:59:37 <phantomcircuit> sipa, without measuring link bandwidth i doubt you can really achieve saturation
12 2015-01-24 01:00:00 <phantomcircuit> since there's no clean way to cancel a request from a slow peer