From: Mohamed M Moustafa (mmma@gawab.com)
Date: Tue Jun 26 2007 - 06:50:51 ART
Hi Edison,
I totally agree with you, by default cisco considers a missing MED as 0,
and "bgp bestpath med missing-med-worst" changes this default behavior by
considering a missing MED as infinite (worst).
But i think that before preferring the route from the router with the
lowest RID, the router prefers the oldest most stable eBGP route.
BR,
Mohammed Mahmoud.
Edison Ortiz <edisonmortiz@gmail.com> wrote on 25 Jun 2007, 11:18 PM:
Subject: Re: BGP Best path selection
>Actually, most documents I've read confirm what you stated,
>the lack of metric attribute on the route defaults to zero.
>
>However, during my labbing, that was not the case. On equal
>routes, the one with a MED value broke the tie-breaker.
>
>Play around and set a MED 0 from both eBGPs. You will see
>the lowest RID will win.
>__
>
>Edison Ortiz
>(Routing & Switching, CCIE # 17943)
>
>
>
>----- Original Message -----
>From: "Bit Gossip" <bit.gossip@chello.nl>
>To: "Edison Ortiz" <edisonmortiz@gmail.com>
>Cc: <ccielab@groupstudy.com>
>Sent: Monday, June 25, 2007 4:07 PM
>Subject: Re: BGP Best path selection
>
>
>> Vvvvvvvvvvery interesting: I was not aware of that. Do you have any
>> reference to the doc for this?
>> So this is the sequence of events:
>> - each of the 2 routers in AS54 originates only one of the prefixes and
>> in doing so attaches MED=0
>> - they pass the prefix to eachother via iBGP preserving MED=0
>> - both of them advertise both prefixes to the eBGP peer but the prefix
>> locally originated maintains MED=0 while for the prefix learned via iBGP
>> the MED is scrapped out
>> - the eBGP peer receives both prefixes from both routers and prefers the
>> one with MED=0 because that one is announced by the router originating
>> it.
>>
>> Is this correct?
>> Thanks,
>> Bit
>>
>>
>>
>> On Sun, 2007-06-24 at 19:20 -0400, Edison Ortiz wrote:
>>> The routes are identical except for the missing MED attribute.
>>>
>>> MED is use as a hint to external neighbors about preferred path into an
>>> AS
>>> that has multiple entry points.
>>>
>>> A route with a MED attribute will be chosen over a route without a MED
>>> attribute
>>> since it's more trustworthy.
>>>
>>> By default, locally generated routes will add MED 0 as one of its
>>> attributes, again
>>> this is to prevent any loops within iBGP.
>>>
>>> If you were to advertise MED with a value of 0 from both routers, the
>one
>>> with the
>>> lowest RID will win.
>>>
>>> HTH,
>>>
>>> __
>>>
>>> Edison Ortiz
>>> (Routing & Switching, CCIE # 17943)
>>>
>>>
>>>
>>> ----- Original Message -----
>>> From: "Bit Gossip" <bit.gossip@chello.nl>
>>> To: "Edison Ortiz" <edisonmortiz@gmail.com>; <ccielab@groupstudy.com>
>>> Sent: Sunday, June 24, 2007 3:03 PM
>>> Subject: Re: BGP Best path selection
>>>
>>>
>>> > This is true; but how can this influence the selction?
>>> > AFAIK: the router running the selection prefers locally originated
>>> > routes
>>> > over learned routes, with locally meaning on the same router
>performing
>>> > the selection.
>>> > In this case none of the prefixes is local and both are learned !
>>> >
>>> > Bit.
>>> >
>>> > ----- Original Message -----
>>> > From: "Edison Ortiz" <edisonmortiz@gmail.com>
>>> > To: "Bit Gossip" <bit.gossip@chello.nl>; <ccielab@groupstudy.com>
>>> > Sent: Sunday, June 24, 2007 8:34 PM
>>> > Subject: Re: BGP Best path selection
>>> >
>>> >
>>> >>I should've read the whole message before replying :)
>>> >>
>>> >> 28.119.16.0/24 was locally originated from 204.12.1.254
>>> >> while
>>> >> 112.0.0.0/8 was locally originated from 54.1.3.254
>>> >>
>>> >> __
>>> >>
>>> >> Edison Ortiz
>>> >> (Routing & Switching, CCIE # 17943)
>>> >>
>>> >>
>>> >>
>>> >> ----- Original Message -----
>>> >> From: "Bit Gossip" <bit.gossip@chello.nl>
>>> >> To: <ccielab@groupstudy.com>
>>> >> Sent: Sunday, June 24, 2007 11:41 AM
>>> >> Subject: BGP Best path selection
>>> >>
>>> >>
>>> >>> maybe I should take a break....
>>> >>> I can not understand the BGP selection process for these 2 prefixes:
>>> >>> -for 28.119.16.0/24 the best next-hop is 204.12.1.254
>>> >>> -for 112.0.0.0 instead it is 54.1.3.254
>>> >>> Any idea why? The difference seems to be that the path with MED=0 is
>>> >>> preferred
>>> >>> over the path with missing MED, but I have NOT configured 'bgp
>>> >>> bestpath
>>> >>> med
>>> >>> missing-as-worst' and so missing MED should be treated as MED=0
>>> >>>
>>> >>> Rack1R6#show ip bgp 28.119.16.0/24
>>> >>> BGP routing table entry for 28.119.16.0/24, version 17
>>> >>> Paths: (2 available, best #2, table Default-IP-Routing-Table)
>>> >>> Flag: 0x820
>>> >>> Advertised to update-groups:
>>> >>> 1 2
>>> >>> 54
>>> >>> 54.1.3.254 from 54.1.3.254 (212.18.3.1)
>>> >>> Origin IGP, localpref 100, valid, external
>>> >>> 54
>>> >>> 204.12.1.254 from 204.12.1.254 (31.3.0.1)
>>> >>> Origin IGP, metric 0, localpref 100, valid, external, best
>>> >>>
>>> >>>
>>> >>>
>>> >>> Rack1R6#show ip bgp 112.0.0.0
>>> >>> BGP routing table entry for 112.0.0.0/8, version 4
>>> >>> Paths: (2 available, best #2, table Default-IP-Routing-Table)
>>> >>> Advertised to update-groups:
>>> >>> 1 2
>>> >>> 54 50 60
>>> >>> 204.12.1.254 from 204.12.1.254 (31.3.0.1)
>>> >>> Origin IGP, localpref 100, valid, external
>>> >>> 54 50 60
>>> >>> 54.1.3.254 from 54.1.3.254 (212.18.3.1)
>>> >>> Origin IGP, metric 0, localpref 100, valid, external, best
>>> >>>
>>> >>>
>_______________________________________________________________________
>>> >>> 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
>
>_______________________________________________________________________
>Subscription information may be found at:
>http://www.groupstudy.com/list/CCIELab.html
>
---------------------------------------------
Free POP3 Email from www.Gawab.com
Sign up NOW and get your account @gawab.com!!
This archive was generated by hypermail 2.1.4 : Sun Jul 01 2007 - 17:24:51 ART