From: Djerk Geurts (djerk@djerk.nl)
Date: Tue Aug 07 2007 - 10:39:41 ART
There's no need. I must say you did make me look...
EIGRP calculates it's metric and afaik there's no strict limit like OSPF
has. OSPF regards 100Mbps as the fastest link. As such if your network has
bigger pipes than that you might want to alter the refBW.
Djerk
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On
> Behalf Of Szarmach, Douglas
> Sent: dinsdag 7 augustus 2007 15:18
> To: Gregory Gombas
> Cc: Group study
> Subject: RE: EIGRP metric calculation demystified
>
> Actually I think I mixed up OSPF and EIGRP here. The OSPF command to
> change the REFBW setting is:
>
> Router ospf x
> auto-cost reference-bandwidth y
>
> However, there does not appear to be a similar EIGRP setting.
>
> Douglas Szarmach
> Senior Network Engineer
> +1 312 648 3797
>
> CME Group
> A CME/Chicago Board of Trade Company
> 20 South Wacker Drive
> Chicago, Illinois 60606
> cmegroup.com
>
> -----Original Message-----
> From: Gregory Gombas [mailto:ggombas@gmail.com]
> Sent: Tuesday, August 07, 2007 4:45 AM
> To: Szarmach, Douglas
> Cc: Group study
> Subject: Re: EIGRP metric calculation demystified
>
> How can you alter the REFBW?
>
> On 8/6/07, Szarmach, Douglas <Douglas.Szarmach@cmegroup.com> wrote:
> > The 10,000,000 is not a fixed number - that's the REFBW setting and
> > might be altered so safer to say REFBW/BW.
> >
> > Also, the interface delay is expressed in '10's of microseconds' in
> > various parts of IOS (stored in units of 10 in the config):
> >
> > Router(config-if)#delay ?
> > <1-16777215> Throughput delay (tens of microseconds)
> >
> > And in regular microseconds in others (including most show
> commands):
> >
> > * x.x.x.x, from y.y.y.y, 1w2d ago, via TenGigabitEthernet1/3
> > Route metric is 29440, traffic share count is 1
> > Total delay is 150 microseconds, minimum bandwidth is
> 100000 Kbit
> > Reliability 253/255, minimum MTU 1500 bytes
> > Loading 8/255, Hops 5
> >
> > --- and the EIGRP formula calls for total delay in just microseconds
> so
> > the 'delay' needs to be adjusted accordingly, many times
> involving the
> > multiplication of the configured interface delay by 10. I don't
> > understand where the 'dividing delay by 10' came from
> though...perhaps
> > working in reverse and returning to the values from running-config?
> >
> > Douglas Szarmach
> > Senior Network Engineer
> > +1 312 648 3797
> >
> > CME Group
> > A CME/Chicago Board of Trade Company
> > 20 South Wacker Drive
> > Chicago, Illinois 60606
> > cmegroup.com
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
> Of
> > Gregory Gombas
> > Sent: Monday, August 06, 2007 10:10 AM
> > To: Group study
> > Subject: EIGRP metric calculation demystified
> >
> > Cisco documentation shows the EIGRP metric calculation as this:
> > metric = [K1*bandwidth + (K2*bandwidth)/(256 - load) + K3*delay] *
> > [K5/(reliability + K4)]
> >
> > Popular literature says it breaks down to this:
> > 256(BW + Delay)
> >
> > In reality the calculation is this:
> > 256((10,000,000/BW)) + (Delay/10))
> >
> > Does anyone know if this final formula can be found on the DOC cd?
> >
> > The closes match I could find was on WARP:
> > http://www.cisco.com/warp/public/103/3.html
> >
> > But can this handy formula be accessed from the lab? Its kind of a
> > pain to memorize.
> >
> > Regards,
> > Greg
> >
> >
> ______________________________________________________________
> _________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
> ______________________________________________________________
> _________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sat Sep 01 2007 - 11:32:09 ART