[Iccrg] review of Compound TCP draft

lachlan.andrew@gmail.com (Lachlan Andrew) Sat, 16 February 2008 00:56 UTC

From: lachlan.andrew@gmail.com
Date: Sat, 16 Feb 2008 00:56:19 +0000
Subject: [Iccrg] review of Compound TCP draft
In-Reply-To: <aa7d2c6d0711211437x4f9befe1ocba11f90fb6019ef@mail.gmail.com>
References: <20071101152220.GB16742@grc.nasa.gov> <FCA794787FDE0D4DBE9FFA11053ECEB60C6CFDFBDE@NA-EXMSG-C110.redmond.corp.microsoft.com> <aa7d2c6d0711211252x2f8660c4w4381eac4990fcf93@mail.gmail.com> <47B05CB1-0F0A-4E57-9DF3-78C0F5A39CB0@nuim.ie> <aa7d2c6d0711211437x4f9befe1ocba11f90fb6019ef@mail.gmail.com>
Message-ID: <aa7d2c6d0802151659y285f648axe9f3c3612011c51a@mail.gmail.com>
X-Date: Sat Feb 16 00:56:19 2008

Greetings Murari,

In the C-TCP specification, it is probably also worth pointing out
that the formula for  dwnd  only applies if  dwnd  is measured in
*packets*, not bytes.

Most RFCs seem to implicitly assume that the variable  cwnd  refers to
the number of bytes (even though it is incremented in multiples of
MSS).  For most flow control, this makes no difference.  However,
    dwnd = 1 packet    =>   dwnd^0.75 = 1
    dwnd = 1500 bytes  =>  dwnd^0.75 = 241
and so increase rule (3) in  draft-sridharan-tcpm-ctcp-01.txt  is much
more aggressive when  dwnd  is measured in packets.

Cheers,
Lachlan

-- 
Lachlan Andrew  Dept of Computer Science, Caltech
1200 E California Blvd, Mail Code 256-80, Pasadena CA 91125, USA
Ph: +1 (626) 395-8820    Fax: +1 (626) 568-3603
http://netlab.caltech.edu/~lachlan

[Iccrg] review of Compound TCP draft Lachlan Andrew
[Iccrg] review of Compound TCP draft weddy@grc.nasa.gov
[Iccrg] review of Compound TCP draft Douglas Leith
[Iccrg] review of Compound TCP draft Lachlan Andrew
[Iccrg] review of Compound TCP draft Murari Sridharan
[Iccrg] review of Compound TCP draft weddy@grc.nasa.gov
[Iccrg] review of Compound TCP draft Lachlan Andrew