Re: Router trick - how to allow only one single packet

I do a fair amount of system test and try to recreate problems that we
see in the field so I am trying to simulate these failure scenarios
that we might see with 4G - oops I said too much already:)

On Thu, Nov 4, 2010 at 11:23 AM, Marko Milivojevic <markom_at_ipexpert.com> wrote:
> If you can find information in that packet that identifies retries,
> you can filter them out using FPM. Otherwise - good luck :-)
>
> Of course, the solution of your problem is to fix the application to
> gracefully handle retries, or communicate properly (if that is
> possible). Not everything is a network problem :-)
>
> --
> Marko Milivojevic - CCIE #18427
> Senior Technical Instructor - IPexpert
>
> FREE CCIE training: http://bit.ly/vLecture
>
> Mailto: markom_at_ipexpert.com
> Telephone: +1.810.326.1444
> Web: http://www.ipexpert.com/
>
> On Thu, Nov 4, 2010 at 15:19, Rich Collins <nilsi2002_at_gmail.com> wrote:
>> There's really very little to distinguish the retries other than some
>> encoded timing information. The client assumes that the packets were
>> dropped or lost.
>>
>> I am tending towards the idea of per packet loading balancing to
>> sinkholes. I just have to deal with about 20 retries.
>>
>> Thnks
>>
>> On Thu, Nov 4, 2010 at 10:54 AM, Marko Milivojevic <markom_at_ipexpert.com> wrote:
>>> If the retries are somehow distinguishable from the original requests,
>>> you may want to consider using FPM.
>>>
>>> --
>>> Marko Milivojevic - CCIE #18427
>>> Senior Technical Instructor - IPexpert
>>>
>>> FREE CCIE training: http://bit.ly/vLecture
>>>
>>> Mailto: markom_at_ipexpert.com
>>> Telephone: +1.810.326.1444
>>> Web: http://www.ipexpert.com/
>>>
>>> On Thu, Nov 4, 2010 at 14:48, Rich Collins <nilsi2002_at_gmail.com> wrote:
>>>> Hi all,
>>>>
>>>> I am trying to test a client application in the lab and need a method
>>>> to block subsequent requests to a server. The retries (UDP packets
>>>> with same length, port number) etc. from this client should not reach
>>>> the server. The retries occur less than a second later and continue.
>>>>
>>>> Limiting by CAR would still pass some of the requests a few seconds
>>>> later. I can't record and spoof this first packet because of the
>>>> encoding in the packet.
>>>>
>>>> I was also thinking of load balancing by packet and creating numerous
>>>> sinkholes at dummy destinations.
>>>>
>>>> Any ideas or EEM?
>>>>
>>>> Thanks
>>>> Rich

Blogs and organic groups at http://www.ccie.net
Received on Thu Nov 04 2010 - 11:39:30 ART