RE: MPLS EIGRP backdoor links - SoO

From: Brian McGahan <bmcgahan_at_ine.com>
Date: Tue, 1 Nov 2011 23:21:30 -0500

SoO is still functional regardless of that command, but it's redundant and not needed because of Cost Community.

Brian McGahan, CCIE #8593 (R&S/SP/Security)
bmcgahan_at_INE.com

Internetwork Expert, Inc.
http://www.INE.com

________________________________________
From: Tom Kacprzynski [tom.kac_at_gmail.com]
Sent: Tuesday, November 01, 2011 10:47 PM
To: Paul Negron
Cc: Brian McGahan; Jochen Bartl; Cisco certification
Subject: Re: MPLS EIGRP backdoor links - SoO

Thank you Christian, Ben, Paul, Jochen and Brian for all of this great information!!!

So based on what Brian mentioned, for us to test EIGRP SoO should we configure the "bgp bestpath cost-community ignore" command on all PE routers running EIGRP to see the SoO functionality during our studies?

Thank you again for these very insightful email responses.

Tom

On Tue, Nov 1, 2011 at 9:56 PM, Paul Negron <negron.paul_at_gmail.com<mailto:negron.paul_at_gmail.com>> wrote:
>
> I too VERY MUCH agree with this statement.
>
> Paul
> --
> Paul Negron
> CCIE# 14856 CCSI# 22752
> Senior Technical Instructor
>
>
>
> > From: Brian McGahan <bmcgahan_at_ine.com<mailto:bmcgahan_at_ine.com>>
> > Reply-To: Brian McGahan <bmcgahan_at_ine.com<mailto:bmcgahan_at_ine.com>>
> > Date: Tue, 1 Nov 2011 21:20:13 -0500
> > To: Jochen Bartl <jochen.bartl_at_gmail.com<mailto:jochen.bartl_at_gmail.com>>
> > Cc: Tom Kacprzynski <tom.kac_at_gmail.com<mailto:tom.kac_at_gmail.com>>, Cisco certification
> > <ccielab_at_groupstudy.com<mailto:ccielab_at_groupstudy.com>>
> > Conversation: MPLS EIGRP backdoor links - SoO
> > Subject: Re: MPLS EIGRP backdoor links - SoO
> >
> > As a side note, EIGRP SoO is a legacy feature that's not needed anymore,
> > because the EIGRP Cost Community encodes the individual vectors of the
> > composite metric calculation into BGP. This way the remote sites run a normal
> > feasibility condition against the end to end path, and it's not possible to
> > cause a loop between sites.
> >
> > SoO isn't needed unless you're running old code that doesn't support Cost
> > Community. It's possible you could get tested on it in the lab but it's not a
> > practical problem anymore.
> >
> > Brian McGahan, CCIE #8593 (R&S/SP/Security)
> > bmcgahan_at_INE.com
> >
> > Internetwork Expert, Inc.
> > http://www.INE.com
> >
> > On Nov 1, 2011, at 5:22 PM, "Jochen Bartl" <jochen.bartl_at_gmail.com<mailto:jochen.bartl_at_gmail.com>> wrote:
> >
> >> Since Paul has already mentioned that you can verify the SoO value
> >> with "show ip eigrp topology XXXX", I would like to throw in some
> >> additional info here. I`ve built a mini lab because I wanted to know
> >> in detail how EIGRP carries the SoO value across the backdoor link.
> >>
> >> After capturing some packets[1] on the backdoor link I could find out
> >> that EIGRP is using the TLV type 260 for SoO. The data inside the TLV
> >> looks something like this.
> >>
> >> 01 04 -> TLV Type (260)
> >> 00 14 -> TLV Length: 20 bytes
> >> 0a 00 19 00 -> TLV references IP routing TLV with route 10.0.25.0
> >> 00 00 00 -> Not sure yet?
> >> 08 -> Ext Community Lengh? 8 bytes
> >> 00 03 -> Ext Community Type?
> >> 00 64 00 00 00 02 -> SOO Ext communit value: 100:2
> >>
> >> I haven't figured out all the header fields yet, but I could at least
> >> identify the SoO value of 100:2 in the last 6 bytes of the TLV. The
> >> hex data is from packet 38 in the capture file[1]. Wireshark/Scapy
> >> doesn't decode the TLV at the moment, but I promise to write a patch
> >> as soon as I've identified all the header fields correctly ;-)
> >>
> >> Petr Lapukhov has also written a great document about EIGRP SoO:
> >>
> http://blog.ine.com/2010/04/29/understanding-eigrp-soo-and-bgp-cost-community>>
> /
> >>
> >> Best regards,
> >>
> >> Jochen
> >>
> >>
> >> 1) http://www.cloudshark.org/captures/327a223fadf3
> >>
> >> On Tue, Nov 1, 2011 at 8:14 PM, Tom Kacprzynski <tom.kac_at_gmail.com<mailto:tom.kac_at_gmail.com>> wrote:
> >>> Hello,
> >>> I'm working on a lab using EIGRP as the protocol for MPLS VPN. This
> >>> scenario has a backdoor link. I'm testing the Site of Origin feature, but
> >>> can't seem to find any way to check the actual SoO on the CE router using
> >>> EIGRP commands. Does anyone know how/if I could do that?
> >>>
> >>> I can see it in BGP table but how do I see it in EIGRP
> >>>
> >>> Rack1R6#sh ip bgp vpnv4 rd 100:1 155.1.67.0/24<http://155.1.67.0/24>
> >>> BGP routing table entry for 100:1:155.1.67.0/24<http://155.1.67.0/24>, version 338
> >>> Paths: (1 available, best #1, table VPN-A)
> >>> Advertised to update-groups:
> >>> 1
> >>> Local
> >>> 0.0.0.0 from 0.0.0.0 (150.1.6.6)
> >>> Origin incomplete, metric 0, localpref 100, weight 32768, valid,
> >>> sourced, best
> >>> Extended Community: SoO:100:60 RT:100:1
> >>> Cost:pre-bestpath:128:28160 <----------------------
> >>> 0x8800:32768:0 0x8801:10:2560 0x8802:65280:25600 0x8803:65281:1500
> >>> mpls labels in/out 22/nolabel(VPN-A)
> >>>
> >>>
> >>>
> >>> Thank you,
> >>>
> >>> Tom
> >>>
> >>>
> >>> 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
> >
> >
> > 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
Received on Tue Nov 01 2011 - 23:21:30 ART

This archive was generated by hypermail 2.2.0 : Thu Dec 01 2011 - 06:29:31 ART