On Thu, Aug 07, 2003 at 11:49:08PM +0100, Gordan wrote: > On Thursday 07 August 2003 19:45, Ian Clarke wrote: > > > Imagine you go to McDonalds and ask a server for some food, they take > > your order. Now, you didn't know, but that server is actually serving > > 20 other people at the same time and consequently it takes you ages to > > get your food. Wouldn't it be better if that server said "Sorry Sir, > > I'm really busy - please try another server". > > Isn't this precisely what limiting simultaneous connections does already? They > are two different things, and I think they should be used in conjunction. > > Byte-by-byte transfers are inefficient, I agree. Would it not be better, then > to do it packet-by-packet instead? It reduces the granularity, so it should > still work OK, even it is marginaly more bursty.
Of course we do it in packets. We have a packet size limit of max(2048, bandwidth limit per second / 3). We also limit trailing field packets (which tend to be large) down to the available bandwidth, if there is some (it is refreshed every 100ms...). And we aggregate messages to send into larger packets (some of our messages are really small), if we can do it without adding latency. > > Gordan -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so.
pgp00000.pgp
Description: PGP signature
