Re: QOS Question

Same theory ...

Lets say each ping packet is 100 bytes. The task says drop 1 out of 10.
To accomplish this, you would set the police configs at police 9xx ... ,
something like this will let through 9 packets, but there will not be enough
for the 10th packet.

My guess is that the task has given you the number of packets and frame
size. The rest is just some simple math ... Also, I have 'rough guessed' it
with the 9xx ..., if a grading script was being used, it would likely look
for the exact numbers.

Nuno, this is how I understand this task. I would be interested to hear
others thoughts if this is out of line ...

Wish you a great night! Happy studying, HTH,

Andrew

On Wed, Dec 30, 2009 at 7:55 PM, Nuno Reis <nreis_at_regra.pt> wrote:

> Hi,
>
> I was not referring to any problem. I'm just trying to solve a challenge I
> saw in a workbook which asks to cause packet loss on icmp. I know it is a
> strange challenge but I was wondering how can I intentionally affect traffic
> without congestion. Sadly I don't have the solution with me.
> I think Andrew got my idea correctly however I'm not sure how to apply it
> here, because is not regarding throughput. I want to drop, has an example, 1
> icmp packet out of 10. But again, the challenge by itself sounds dummy,
> but... is it doable?
>
> Cheers,
> Nuno
> -----Original Message-----
> From: ben edgar [mailto:bedgar2500_at_gmail.com]
> Sent: Wednesday, December 30, 2009 8:49 PM
> To: ALL From_NJ
> Cc: Nuno Reis; ccielab_at_groupstudy.com
> Subject: Re: QOS Question
>
> what type of router is it? how many PPS are being sent?
> sometimes the router can be overloaded in other ways other then pure
> bit's per second.
>
> also how are you measuring the amount of traffic being sent/congestion?
> depending on the type of traffic flow you may actually have congestion
> but on the 30 second or by default 5 minute measuring interval on a
> Cisco router
> you may not see it as congestion.
>
> Ben
>
> On Wed, Dec 30, 2009 at 11:58 AM, ALL From_NJ <all.from.nj_at_gmail.com>
> wrote:
> > Hello,
> >
> > I hope this finds you doing well.
> >
> > Interesting question ... humm. If there is no congestion, then packets
> are
> > sent directly to the out tx ring and as such, no drops.
> >
> > Two follow ups to this ...
> >
> > If you want to drop traffic over a specific throughput, then configure
> > policing. You can drop traffic that exceeds your configured amount, or
> that
> > exceeds any amount you prefer. Get out the calculator Nuno, and figure
> out
> > what 20% of the total traffic is, and then configure your police rate at
> 20%
> > lower than what you are sending. Make sense? You can police at 20%
> lower
> > ...
> >
> > Another way to achieve queuing when there is no interface congestion, is
> to
> > use shaping. Shaping will create some delay to the traffic flows and use
> > buffers / software queuing. This is not really related to your question
> > about drops, however when you need to send at a rate less than your
> > interface, then shaping is your best option.
> >
> > HTH,
> >
> > Andrew Lee Lissitz
> >
> >
> >
> >
> > On Wed, Dec 30, 2009 at 9:29 AM, Nuno Reis <nreis_at_regra.pt> wrote:
> >
> >> Hi all,
> >>
> >> Is there any way of causing packet drops using qos mechanism, i.e
> causing
> >> 20% packet loss even if there is no congestion. WRED?
> >>
> >> Thanks,
> >> Nuno Reis
> >>
> >>
> >> Blogs and organic groups at http://www.ccie.net
> >>
> >> _______________________________________________________________________
> >> Subscription information may be found at:
> >> http://www.groupstudy.com/list/CCIELab.html
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >
> >
> > --
> > Andrew Lee Lissitz
> > all.from.nj_at_gmail.com
> >
> >
> > Blogs and organic groups at http://www.ccie.net
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>

-- 
Andrew Lee Lissitz
all.from.nj_at_gmail.com
Blogs and organic groups at http://www.ccie.net