All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jody Shumaker <jody.shumaker@gmail.com>
To: lartc@vger.kernel.org
Subject: Re: [LARTC] external and internal LANs
Date: Wed, 25 May 2005 23:55:43 +0000	[thread overview]
Message-ID: <4295107F.80005@gmail.com> (raw)
In-Reply-To: <4074fa403a74.403a744074fa@idm.net.lb>

What basically happens is that the slowest link ends up controlling the 
shaping.  Usually this is the device connecting you to your ISP, and it 
tends to have a very simple buffering of data, FIFO, first in first 
out.  What this ends up meaning is that when you send data to it at the 
much faster rate of your local network, eventually you fill the buffer, 
at that point it just starts dropping packets. It's actually much more 
complicated than this, but all you need to know is that he buffer before 
the slowest link can only hold so many,  and how it sends out what its 
holding is the important thing.  With a simple fifo queue you may have a 
bunch of data from an ftp transfer sitting in the queue, then you add a 
http request to the end of the queue.  The http request has to wait till 
all that ftp data gets sent, even though it's a really tiny bit of data.

If you instead shape the data on the linux router, you can make yourself 
the slowest link.  By doing so you prevent the simple queue from filling 
up much, and can implement some sort of priorities for traffic.  If you 
gave http request more priority than ftp data, then even if your queue 
is full of ftp data, when the http request comes along it will get sent 
as soon as possible instead of waiting for all the ftp packets already 
on the queue to be sent out.

Giving TCP Ack packets higher priority is another common practice. Even 
when you're just downloading data, from say a website or ftp, ackowledge 
packets are sent to ackowledge you got each piece of data.  The other 
end will only send so much data before it waits to receive ack packets.  
Giving these rather small packets priority, your downloads will never 
suffer because of your uploads.



rani79@idm.net.lb wrote:

>hi all
>what does it mean if the total network speed on internal LAN is greater 
>than the external LAN 
>and why it happens?
>the extrnal LAN here is the Intenet using a dedicated bandwidth. and 
>BTW the linux box NATting
>has no shapping at all. just direct piping
>
>_______________________________________________
>LARTC mailing list
>LARTC@mailman.ds9a.nl
>http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
>
>  
>

_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

  reply	other threads:[~2005-05-25 23:55 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-05-25 10:52 [LARTC] external and internal LANs rani79
2005-05-25 23:55 ` Jody Shumaker [this message]
2005-05-26 11:57 ` rani79

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4295107F.80005@gmail.com \
    --to=jody.shumaker@gmail.com \
    --cc=lartc@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.