Re: Removing T/TCP and replacing it with something simpler

From: Andre Oppermann (andre_at_freebsd.org)
Date: 10/22/04

  • Next message: Mark Allman: "Re: Removing T/TCP and replacing it with something simpler" 
    Date: Fri, 22 Oct 2004 17:55:45 +0200
To: Brian Fundakowski Feldman <green@freebsd.org>

    Brian Fundakowski Feldman wrote:
    >
    > On Fri, Oct 22, 2004 at 05:14:07PM +0200, Andre Oppermann wrote:
    > > None of it. Neither NOPUSH nor CORK have any security implications.
    > > Those are only with the specification of T/TCP. Blocking the data
    > > is independend of 3WSH. Normally you have Nagle enabled (default)
    > > and when you don't fill an entire packet worth of data it will wait
    > > up to 200ms to send the packet in anticipation of more data from the
    > > socket. This screws the responsiveness of your connection. The first
    > > solution is to turn off Nagle (with TCP_NODELAY) but now you get a
    > > packet for every single write() you do. Fine for telnet and ssh but
    > > not the right thing for a database server. There you don't want the
    > > delay but at the same time you want several successive write()s that
    > > will go in one packet on the wire. Here NOPUSH and CORK come into
    > > play.
    >
    > Why is just tuning the delay a bad solution?

    If you tune it too low it ain't useful anymore (doesn't gather distant
    writes together) and too many timers too often. 

    -- 
Andre
_______________________________________________
freebsd-arch@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-arch
To unsubscribe, send any mail to "freebsd-arch-unsubscribe@freebsd.org"
  • Next message: Mark Allman: "Re: Removing T/TCP and replacing it with something simpler"

    Relevant Pages

    • Re: Should i disable Nagels Algorithm in my http client?
      ... The data in the buffer is packaged and sent immediately UNLESS ... stream a packet will be sent. ... turn off Delayed Acknowledgements or adjust the delay time. ... The delayed ACK of the second packet doesn't have any effect because the ...
      (comp.lang.perl.misc)
    • Re: how much size is enough ?
      ... If I assume my minimum rx packet length of 64 bytes and maximum packet ... You are not going to be able to do 200 metre segments with existing ... and available fibre types) affects the bit encoding used, ... The 802.3ae-2002 standard indicates that a conservative delay estimate ...
      (comp.dcom.lans.ethernet)
    • Re: estimate the cdf 95% with a confidence interval of a 95%
      ... in each file I have the delay of all the packets that the station has ... then if packet n saw a delay very high because the ... to again compute the 95% cdf out of them. ... distribution function for a binomial distribution, ...
      (sci.math)
    • Re: Bandwidth
      ... >simulate the network at a bandwidth of 9.6kbps). ... The time from when you START sending the packet to the time when the ... The speed-of-light propagation delay is negligible ... in twisted-pair or shielded cables is somewhat less than the speed ...
      (comp.unix.programmer)
    • Re: Socket Send
      ... without delay, but without TCP_NODELAY a delay is almost guaranteed in this ... If you have no idea what the timing pattern of your data is going to ... know better than the TCP stack how to segment the data. ... packet, so there's no problem with acknowledged data that doesn't ...
      (comp.unix.programmer)