home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.protocols.tcp-ip.ibmpc:4745 news.software.readers:1656
- Newsgroups: comp.protocols.tcp-ip.ibmpc,news.software.readers
- Path: sparky!uunet!wupost!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!odin!trier
- From: trier@odin.ins.cwru.edu (Stephen C. Trier)
- Subject: Re: Trumpet Shell for FTP Software TCP stack.
- Message-ID: <1992Aug16.032047.11419@usenet.ins.cwru.edu>
- Sender: news@usenet.ins.cwru.edu
- Nntp-Posting-Host: odin.ins.cwru.edu
- Organization: Case Western Reserve University, Cleveland, OH (USA)
- References: <kemp.713806943@convex.convex.com> <1992Aug15.180822.26989@news.csuohio.edu>
- Date: Sun, 16 Aug 92 03:20:47 GMT
- Lines: 21
-
- In article <1992Aug15.180822.26989@news.csuohio.edu> cowboy@trans.csuohio.edu (Syscrusher) writes:
- >Try reducing your mss to mss=256 and reduce your rwin to rwin=256. This
- >might help, particularly with a SLIP connection.
-
- This is not always the best advice. The mss should definitely be in the
- 200 range, but the receive window can stay large. Making the receive window
- as short as the MSS makes TCP very inefficient.
-
- MSS-sized windows cause TCP to degenerate into a protocol more like XMODEM
- in which every packet must be acknowledged before the next can be sent. It
- can't stream the data out like it can when the receive window is a reasonable
- size. (In the latter case, the sender can start transmitting the next data
- packet *while waiting for* an acknowledgement for the last. When everything
- goes right, there are no sit-and-wait-for-a-reply delays.)
-
- --
- Stephen Trier "I am not a lawyer, though I play one
- Network Services Engineering on the net."
- CWRU IRIS/INS/Telecom Chris Torek, torek@horse.ee.lbl.gov
- trier@ins.cwru.edu
-