All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sorin Panca <sorin.panca@gmail.com>
To: netfilter@lists.netfilter.org
Subject: Re: How would this help my LAN / network
Date: Fri, 07 Oct 2005 20:36:33 +0300	[thread overview]
Message-ID: <4346B221.8010205@gmail.com> (raw)
In-Reply-To: <1128705132.5537.10.camel@localhost>

Hi John!

John A. Sullivan III wrote:
> Unless one has contracted with a private network, one is not likely to
> find QoS/CoS.  However, your bottleneck is likely to be the last mile.
> Although the ISP router may not be regulating the QoS on the last mile,
> you can control the ingress and egress through your gateway . . . well,
> at least the egress and the processing of the ingress - John
> 
> On Fri, 2005-10-07 at 11:45 -0500, Gregory G Carter wrote:
> 
I think TOS is for admins who want to priorize _THEIR_ traffic in
_THEIR_ controlled networks. ISP don't do any sahping. They only have
large txqueues and with them they demonstrate how fast the download is.
If I show them that when i download with the maximum speed and I cannot
browse the internet or chat, they say to stop downloading so fast.
When a tiny little http response packet waits in a 2 MBytes queue to
leave their server to my computer with a link speed of 128 kbit/s, I
have to wait a very long time for it to arrive and meanwhile the browser
times out...
So LARTC people developed ingress policing, to drop the bulk packets and
to sinal this loss to the isp's router, who discards a lot of the
traffic in his queue so my tiny little packet gets in front of other
packets and arrives to me in time...
Wouldn't it be better if ISPs would do some traffic shaping? (No more
packet loss, less internet traffic (discarded packets are resent at some
point), and better browsing... and... happy costomers.
>>Upstream providers such as my ISP here (Time Warner Cable) hardly 
>>support ECN, I would be surprised if they did.
>>
>>For example, the crappy Zyxel cable modems they put out here don't 
>>support ECN notification, so the first thing they do when they overflow 
>>due to high amounts of traffic is simple shut down, most of them just crash.
>>
>>I have to put rate limiting policies on all my routers connected to 
>>these modems otherwise they simply die when you try and push too much 
>>data through them.
>>
>>I remember when Time Warner was using ubr9xx routers from Cisco, which 
>>were excellent, but much more expensive of course than the Zyxel.
>>
>>If somone like Time Warner can't do ECN, I would hardly expect them to 
>>do traffic prioritization....Oooo....complicated.
>>
>>:-)
>>
>>-gc
>>
>>Rob Sterenborg wrote:
>>
>>
>>>>>I just came across one of the threads on the LARTC.
>>>>>In the thread it had this.
>>>>>
>>>>>So my question is, for the following rules, would these increase my
>>>>>browsing / traffic, and if so, how.
>>>>>
>>>>>     
>>>>>
>>>>
>>>>Small packets and control pachets should have priority
>>>>because they carry intrractive traffic.
>>>>   
>>>>
>>>
>>>Is it true that the upstream (ISP) routers have to support TOS for this
>>>to work ?
>>>
>>>
>>>Gr,
>>>Rob
>>>
>>>
>>> 
>>>
>>
> 


  reply	other threads:[~2005-10-07 17:36 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-10-07 14:30 How would this help my LAN / network Brent Clark
2005-10-07 15:16 ` Sorin Panca
2005-10-07 16:08   ` Brent Clark
2005-10-07 16:21   ` Rob Sterenborg
2005-10-07 16:40     ` Sorin Panca
2005-10-07 16:45     ` Gregory G Carter
2005-10-07 17:12       ` John A. Sullivan III
2005-10-07 17:36         ` Sorin Panca [this message]
2005-10-07 16:36   ` Brent Clark
2005-10-07 17:21     ` Sorin Panca

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=4346B221.8010205@gmail.com \
    --to=sorin.panca@gmail.com \
    --cc=netfilter@lists.netfilter.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.