On Tuesday 11 March 2008 15:56, Robert Hailey wrote: > Input Rate: 24.5 KiB/sec (of 781 KiB) > Output Rate: 39.7 KiB/sec (of 30.0 KiB) > Total Input: 1.60 GiB (21.4 KiB/sec) > Total Output: 2.39 GiB (32.1 KiB/sec) > Payload Output: 1.10 GiB (14.7 KiB/sec)(46%) > Request output (excluding payload): CHK 225 MiB SSK 423 MiB. > Insert output (excluding payload): CHK 65.8 MiB SSK 7.01 MiB. > Offered keys: sending keys 16.2 MiB, sending offers 7.16 MiB > Swapping Output: 16.7 MiB. > Connection setup: 4.59 MiB output > Ack-only packets: 377 MiB > Resent bytes: 187 MiB > Updater Output: 0 B > Announcement output: 1.05 MiB > Admin bytes: 77.6 KiB initial messages, 4.06 KiB IP change messages, > 9.42 KiB disconnection notifications, 0 B routing status > Debugging bytes: 0 B network coloring, 0 B ping, 26.1 KiB probe > requests, 0 B routed test messages. > Node to node messages: 67.2 KiB > Other output: -8375632 B (0%) > > --build #1122 (+/-)-- > > I like the added stats, but... I've got a question. 16% of my total > output bandwidth is ack-only packets. Isn't that kinda high?
Yes. It sucks. As does the fact that we send a lot of very small packets, which are then padded with random data and subjected to our huge per-packet overhead. Most obvious solutions to this are sufficiently large as to not be possible before 0.7.0. > > In fact, even 8% for resends seems kinda high too, but I suppose that > does not mean 8% packet loss. The negative 'other' output is amusing > but not very useful ;) Mine is positive at this moment in time. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080311/0e83b5df/attachment.pgp>
