Re: How to Quicken TCP Re-transmission?
- From: Mark Allman <mallman@xxxxxxxx>
- Date: Tue, 23 May 2006 09:07:45 -0400
1. Receiver should tell sender to re-send as soon as possible.
(But TCP makes receiver purely passive)
This isn't really going to help you at all. With SACK (especially, but
even without it) the receiver isn't really in a whole lot better
position than the sender to judge when a packet is actually lost. Some
people have worked on SNACKs (selective NEGATIVE acknowledgments), but
my opinion is that the results (that I have seen) show them to be fairly
equivalent to SACK in terms of performance.
2. Receiver should tell sender what is really necessary to re-send.
(Sometimes only a single ACK number of TCP cannot include enough
information)
RFC2018. (Which provides more than a single ACK number. But, this
doesn't make the receiver tell the sender what to resend. The logic
still resides at the sender.)
allman
Attachment:
pgpY9leyAigkx.pgp
Description: PGP signature
- Follow-Ups:
- Re: How to Quicken TCP Re-transmission?
- From: mag
- Re: How to Quicken TCP Re-transmission?
- Prev by Date: Re: improving transport over lossy links ?
- Next by Date: Re: nfsd and CPU/performance problem
- Previous by thread: Re: How to Quicken TCP Re-transmission?
- Next by thread: Re: How to Quicken TCP Re-transmission?
- Index(es):
Relevant Pages
|
|