Re: TCP question

From: Charlie Younghusband ([email protected])
Date: Tue Aug 28 2001 - 03:10:38 EDT

  • Next message: Sergey Raber: "Re: TCP question"

    Argh, I just realized that some of my email is bogus. I was thinking that
    the MSS was much larger than what you using (or even your 64K trick), not simply
    roughly 10x. At that point it becomes more an issue of using a larger but still
    reasonable granularity chunk. It would still scare people running it
    onto a hybrid network such as the greater Internet or even busy LAN
    (which has a surprisingly terrible effect on your sat bandwidth from my
    experience) on one end for classical reasons as already mentioned. But
    for a private network satellite hookup, it's more viable as a tuning
    option given the bandwidth you're playing with. Interesting...

    Charlie

    Charlie Younghusband wrote:

      It does not surprise me that you received performance gains. Looking
    at
      the simple single connection over an uncongested direct satellite link
    as
      part of another project I worked on, we did a simple mathematical
    formula
      where we asked how low does the BER have to go before the largest (RFC
      compliant) TCP MSS size available was no longer optimal, and factoring
    in
      header overhead the answer was ridiculously low (like 10e-3). By
      extension, it was clear that a larger MSS was much more efficient on the
      higher bandwidth links seen today. I'm not sure where you got the 16208
    MSS
      size, but I suspect that it is link bandwidth specific and if for some
      reason your bandwidth was seriously suddenly reduced (such as a
    competing
      data stream) you'd suddenly have virtually no bandwidth available to
    either
      data stream as neither would complete a TCP segment very often. Still,
    it
      is an interesting tuning idea for a single TCP connection over a
    satellite
      link with no other competition when the bandwidth is known and
    fixed. Other
      than that case, things fall apart quickly as there is poor scaling and
      little adaption. As Gorry Fairhurst pointed out, run it with anything
    else
      and then see. There are other better options. (I'll also point out that
      you're still only at best about 73% link efficiency with your
    modifications,
      so even if this specific case always applies to your usage, wasting over
      2Mbps of satellite bandwidth probably won't please whoever pays for it
    :))

      Cheers,
      Charlie

      ---
      Charlie Younghusband
      Network Software Engineering
      Xiphos Technologies http://www.xiphos.ca/
      514-848-9640



    This archive was generated by hypermail 2b29 : Tue Aug 28 2001 - 03:27:25 EDT