Re: Internet Traffic load balancing

From: Marko Milivojevic <markom_at_ipexpert.com>
Date: Mon, 2 May 2011 16:56:05 -0700

Yup - found that one, but that's still 2nd hand information and not
supported by measurements. I asked him directly about it to see if he
tested this with traffic.

I still remain open to the idea that this may work, but I'd like to
see it. I don't have measurement equipment required to test this in
detail.

I wonder if we could test this with a simple ping?

--
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 Mon, May 2, 2011 at 16:53, Carlos G Mendioroz <tron_at_huapi.ba.ar> wrote:
> I don't recall the place where I see it first,
> Ivan Pepelnjak has a nice blog post at
> http://blog.ioshints.info/2006/10/cef-load-sharing-details.html
>
> -Carlos
>
> Marko Milivojevic @ 02/05/2011 20:40 -0300 dixit:
>>
>> I don't follow every topic on GS in detail, but as I said - I haven't
>> tested this.
>>
>> Do you have a link to a document where this is detailed?
>>
>> --
>> 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 Mon, May 2, 2011 at 16:32, Carlos G Mendioroz <tron_at_huapi.ba.ar> wrote:
>>>
>>> AFAIK, it's been documented that it works as Brian says, i.e., 5:2
>>> relation,
>>> or even 7:3 if need be.
>>> This same topic has been here some 2 or 3 months ago ?
>>> -Carlos
>>>
>>> Marko Milivojevic @ 02/05/2011 20:06 -0300 dixit:
>>>>
>>>> To make my argument further:
>>>>
>>>> R2#sh ip cef 0.0.0.0 0.0.0.0 detail
>>>> 0.0.0.0/0, epoch 0, per-destination sharing
>>>> B DefNet source: 0.0.0.0/0
>>>> B recursive via 169.254.1.1
>>>> B  recursive via 10.0.0.1
>>>> B  B  attached to GigabitEthernet0/0
>>>> B recursive via 169.254.1.2
>>>> B  recursive via 10.0.0.1
>>>> B  B  attached to GigabitEthernet0/0
>>>> B recursive via 169.254.1.3
>>>> B  recursive via 10.0.0.1
>>>> B  B  attached to GigabitEthernet0/0
>>>> B recursive via 169.254.1.4
>>>> B  recursive via 10.0.0.1
>>>> B  B  attached to GigabitEthernet0/0
>>>> B recursive via 169.254.1.5
>>>> B  recursive via 10.0.0.1
>>>> B  B  attached to GigabitEthernet0/0
>>>> B recursive via 169.254.2.1
>>>> B  recursive via 20.0.0.2
>>>> B  B  attached to GigabitEthernet0/1
>>>> B recursive via 169.254.2.2
>>>> B  recursive via 20.0.0.2
>>>> B  B  attached to GigabitEthernet0/1
>>>>
>>>> As you can see, CEF resolves this in the manner I described.
>>>>
>>>> --
>>>> 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 Mon, May 2, 2011 at 15:55, Marko Milivojevic <markom_at_ipexpert.com>
>>>> wrote:
>>>>>
>>>>> On Mon, May 2, 2011 at 15:52, Brian McGahan <bmcgahan_at_ine.com> wrote:
>>>>>>
>>>>>> Yes each route is equal, but more point to one exit point vs. another.
>>>>>> Try it, it works.
>>>>>
>>>>> Just to make it clear - I'm not arguing. I like the solution, but I'm
>>>>> not sure if it's going to solve the problem for the reasons below.
>>>>>
>>>>> I'm sure they will show in the routing table as 7 different routes.
>>>>> However, when CEF resolves adjacencies, there will still be only two
>>>>> exits. While we may have the impression that we have 5:2 ratio by
>>>>> looking into the RIB, are we _actually_ going to have it with the
>>>>> traffic.
>>>>>
>>>>> Without generating the traffic and measuring the output, I'm not quite
>>>>> sure how to test this...
>>>>>
>>>>> --
>>>>> 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/
>>>>
>>>> Blogs and organic groups at http://www.ccie.net
>>>>
>>>> _______________________________________________________________________
>>>> Subscription information may be found at:
>>>> http://www.groupstudy.com/list/CCIELab.html
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>> --
>>> Carlos G Mendioroz B <tron_at_huapi.ba.ar> B LW7 EQI B Argentina
>>>
>
> --
> Carlos G Mendioroz B <tron_at_huapi.ba.ar> B LW7 EQI B Argentina
Blogs and organic groups at http://www.ccie.net
Received on Mon May 02 2011 - 16:56:05 ART

This archive was generated by hypermail 2.2.0 : Wed Jun 01 2011 - 09:01:11 ART