Re: EIGRP K values

Nobody pays attention to the formulas, statistically speaking :)
Cisco has, for quite a long time, copy & pasted a phrase that said that
EIGRP metric had MTU involved in the computation. Not so, although EIGRP
does carry MTU as part of the path data, and it could have used it...

Does anybody know of anyone using Ks for dynamic load balancing based on
anything that is not the default ?

-Carlos

Tyson Scott @ 7/10/2010 10:42 -0300 dixit:
> This is the formula that is used by default since K1 and K3 are 1 by default
> and the rest are 0.
> EIGRP Metric = 256*((10^7 / min. Bw) + Delay in micro sec/10)
>
> Take a look at the following blog articles if you want to know more.
> http://blog.ipexpert.com/2010/03/03/eigrp-metric-k-values/
> http://blog.ipexpert.com/2010/05/03/eigrp-unequal-cost-load-balancing/
>
> The first one is more relevant to this topic
>
> Regards,
>
> Tyson Scott - CCIE #13513 R&S, Security, and SP
> Managing Partner / Sr. Instructor - IPexpert, Inc.
> Mailto: tscott_at_ipexpert.com
>
>
> -----Original Message-----
> From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of
> Narbik Kocharians
> Sent: Thursday, October 07, 2010 9:27 AM
> To: Piotr Malarski
> Cc: Cisco certification
> Subject: Re: EIGRP K values
>
> Hahahaha
>
> Both links are from Cisco. Amazing, i did not pay attention to that.
>
> But Both formulas will result to the same composite metric unless the K5
> value is manipulated. Now that i look at both of them a little more
> carefully, i would go with your link. I think they have a typo in the link
> that i provided.
>
> On Thu, Oct 7, 2010 at 6:12 AM, Piotr Malarski
> <piotr.malarski.99_at_gmail.com>wrote:
>
>> The formula from the link that you provided is not the same as the
>> formula from the DOC CD available during lab.
>>
>> #1 : Formula from your link:
>> metric = [K1 * bandwidth + (K2 * bandwidth) / (256 - load) + K3 *
>> delay] * [K5 / (reliability + K4)]
>>
>> #2 : Formula from DOC CD, IOS 12.4 Configuration guide
>> metric = 256*((K1*Bw) + (K2*Bw)/(256-Load) + (K3*Delay)*(K5/(Reliability +
>> K4)))
>>
>> The differences are:
>> a. Compare how K values are grouped with brackets in both formulas. If
>> K5=1 then calculation is different in formula #1 and #2
>> b. In #2 there is multiplication by 256, not in #1
>>
>> One of them (or both) must be wrong. Which one ?
>>
>> I am referring to this DOC CD article:
>>
>>
> http://www.cisco.com/en/US/docs/ios/iproute_eigrp/configuration/guide/ire_cf
> g_eigrp_ps6350_TSD_Products_Configuration_Guide_Chapter.html
>>
>> On Thu, Oct 7, 2010 at 4:44 AM, Narbik Kocharians <narbikk_at_gmail.com>
>> wrote:
>>> If K5 is set to it's default value (Which is zero), the last portion of
>> the
>> > formula (K5/K4+rel) will not come into play:
>>>
> http://www.cisco.com/en/US/tech/tk365/technologies_white_paper09186a0080094c
> b7.shtml
>>>
>>>
>>> On Thu, Oct 7, 2010 at 1:09 AM, Elias Chari <elias.chari_at_gmail.com>
>> wrote:
>>>> Yes, the maths do not work out when you look at it that way. May be
>>>> another
>>>> way to represent the formula, which may make it easier to interpret
>> would
>>>> be:
>>>>
>>>> (a) For K5>0, the formula applied by the router is:
>>>>
>>>> EIGRP Metric = 256*((K1*Bw) + (K2*Bw)/(256-Load) +
>>>> (K3*Delay)*(K5/(Reliability + K4)))
>>>>
>>>> (b) For K5=0, the formula applied by the router is:
>>>>
>>>> EIGRP Metric = 256*((K1*Bw) + (K2*Bw)/(256-Load) + (K3*Delay))
>>>>
>>>>
>>>> On 7 October 2010 07:53, Narbik Kocharians <narbikk_at_gmail.com> wrote:
>>>>
>>>>> If K5 is set to zero (Which is the default), then, the last section
> of
>>>> the
>>>>> formula (K5/(Reliability + K4) will not come into play.
>>>>>
>>>>> On Wed, Oct 6, 2010 at 9:18 PM, Piotr Malarski
>>>>> <piotr.malarski.99_at_gmail.com>wrote:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> What should be K values so only delay is used to calculate the
>> metric ?
>>>>>> OK, I know that the answer is 0 0 1 0 0 (K1,K2,K3,K4,K5) but this
>> does
>>>>>> not fit the Doc CD formula:
>>>>>>
>>>>>> EIGRP Metric = 256*((K1*Bw) + (K2*Bw)/(256-Load) +
>>>>>> (K3*Delay)*(K5/(Reliability + K4)))
>>>>>>
>>>>>> If K3=1 and K4=K5=0 then the last section of the above formula
>> equals
>>>>>> 0 and Delay will not be considered at all:
>>>>>>
>>>>>> (1*Delay)*(0/(Reliability + 0)) = (1*Delay)*0 = 0
>>>>>>
>>>>>> Actually, if K values are 00100 then the whole metric equals zero.
> I
>>>>>> thinks that my math is correct. Can someone please clarify ?
>>>>>>
>>>>>>
>>>>>> Blogs and organic groups at http://www.ccie.net
>>>>>>
>>>>>>
>> _______________________________________________________________________
>>>>>> Subscription information may be found at:
>>>>>> http://www.groupstudy.com/list/CCIELab.html
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> Narbik Kocharians
>>>>> CCSI#30832, CCIE# 12410 (R&S, SP, Security)
>>>>> www.MicronicsTraining.com <http://www.micronicstraining.com/> <
>> http://www.micronicstraining.com/> <
>>>> http://www.micronicstraining.com/>
>>>>> Sr. Technical Instructor
>>>>> YES! We take Cisco Learning Credits!
>>>>> Training And Remote Racks available
>>>>>
>>>>>
>>>> > 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
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>> --
>>> Narbik Kocharians
>>> CCSI#30832, CCIE# 12410 (R&S, SP, Security)
>>> www.MicronicsTraining.com <http://www.micronicstraining.com/>
>>> Sr. Technical Instructor
>>> YES! We take Cisco Learning Credits!
>>> Training And Remote Racks available
>>>
>>>
>>> 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
>>
>>
>>
>>
>>
>>
>>
>>
>
>

-- 
Carlos G Mendioroz  <tron_at_huapi.ba.ar>  LW7 EQI  Argentina
Blogs and organic groups at http://www.ccie.net