http://amphibian.dyndns.org/load.png

Graph of the "mean ping time" on my node this weekend. Actually the
"mean ping time" isn't really a ping time. What it is is the average,
for all connected nodes, of the maximum of two things:
- The average round-trip time calculated from packet acknowledgements
  (after today's commits, included in the above), or the ping time for
  an explicit FNPLinkPing.
- The delay between a data packet being queued to be sent and its
  actually being sent (an indication of how much data we want to send
  but can't because of bandwidth limiting).

This probably is misnamed and is not the best way to measure load, but
in any case this is what we do with it:
- If it is below 1000ms, we accept all requests.
- If it is above 2000ms, we reject (almost) all requests.
- If it is between we accept some requests but not all.

The reason we do this is to prevent timeouts; a timed-out request is
much more costly than one which is just rejected for reasons of
overload.
-- 
Matthew J Toseland - [EMAIL PROTECTED]
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.

Attachment: signature.asc
Description: Digital signature

_______________________________________________
Devl mailing list
[email protected]
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to