On Wed, Apr 12, 2006 at 05:26:51PM +0100, Michael Rogers wrote: > > > When a packet is successfully transmitted we increment the window size > > by b. When a packet is lost we multiply it by a. We send packets every > > (round trip time / window size). > > > > This is definitely TCP-friendly? > > This is definitely not TCP-friendly - it's not even AIMD. You never > leave slow-start mode, so the window size increases exponentially until > you lose a packet.
ftp://ftp.ee.lbl.gov/papers/congavoid.ps.Z footnote on page 6: Increment the window size by 1/window size when a packet succeeds. Halve it when one fails. To adjust it to our usage: When a packet succeeds, winSize += b / winSize When a packet fails, winSize *= a Now, is *that* TCP-compatible? > > In the first RTT there will be one packet and one ack, so the window > will increase by b to 1+b. In the second RTT there will be 1+b packets > and 1+b acks, so the window size will increase by 1+b to 2+2b. In the > third RTT there will be 2+2b packets and 2+2b acks, so the window size > will increase by 2+2b to 4+4b, etc. > > Cheers, > Michael -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20060412/7441d944/attachment.pgp>
