The MPLS-OPS Archive

Cell Relay Retreat>MPLS-OPS Archive>month:2003-May> msg00121



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

RE: OSPF and RIP

  • From: "Lars Higham" <lhigham@yahoo.com>
  • Date: Sat, 17 May 2003 19:00:25 +0530
  • Cc: <mpls-ops@mplsrc.com>
  • Importance: Normal
  • Resent-Date: Sat, 17 May 2003 10:19:25 -0400
  • To: "'Spice Sylvia'" <falsesylvia@yahoo.co.uk>, <alok.dube@apara.com>

Title: Message
Hello Sylvia,
 
You're thinking of the PE-CE routing, where I believe that Alok's concern is that once the routes follow the path CE-PE-MPGBP-PE he doesn't then want to put the routes into the destination PE vrf's IGP, whatever it is.
 
The cleanest way of accomplishing this, depending on how Internet connections and routing are handled by the customer, is to advertise a default route, or even one or more summary routes, into the destination PE's vrf IGP enabling default routing for that site and removing the requirement for installing specific route information into the IGP.
 
Regards,
Lars Higham
-----Original Message-----
From: Spice Sylvia [mailto:falsesylvia@yahoo.co.uk]
Sent: Saturday, May 17, 2003 6:53 PM
To: alok.dube@apara.com; lhigham@yahoo.com
Cc: mpls-ops@mplsrc.com
Subject: RE: [MPLS-OPS]: OSPF and RIP

Hi Alok,
 
You could see if its possible to run OSPF "per VR" and see if that helps.
 
like:
 
R1---VRF1(OSPF)---L2LSPbetween2VRs----VRF2(OSPF)--R2
 
VRF1 and VRF2 can for adjacencies.
 
If LSPs are VRF to VRF, it might be possible to do this.
 
Myself never tried it, perhaps others on the list may have.
 
-S.F.

Alok Dube <alok.dube@apara.com> wrote:
hi Lars,

I am sorry I think I have not worded my question properly.

What I want to know is this

CE1 and cE2 are connected at 2 points in a VPN

they run OSPF/RIP which needs to be distributed into BGP etc and then back
on the other side into RIP...etc..Am i right?
I want to know if one can avoid the "BGP into IGP" part.

Now is there a "multihop" variation of RIP or OSPF that could be run
directly between the CEs?
In other words, the PE needs not put the BGP routes into IGP? but PE needs
to put "IGP into BGP" just for reachability.
am I a bit clearer? or maybe not? :o)










> Alok, see below:
>
> Regards,
> Lars Higham
>
>
> -----Original Message-----
> From: alok [mailto:alok.dube@apara.com]
> Sent: Thursday, May 15, 2003 10:35 PM
> To: mpls-ops@mplsrc.com
> Subject: [MPLS-OPS]: OSPF and RIP
>
>
> Hi,
>
> what distinguishes OSPF from RIP in terms of a IGP in an MPLS network?
> OSPF is more resource intensive
> OSPF has faster convergence
> RIP is generally easier to configure and troubleshoot - lower
> complexity RIP v1 doesn't support VLSM; RIP v2 and OSPF do
> TTBOMK, RIP doesn't support traffic-engineering as do OSPF and ISIS
>
> are protocols in UDP not route-able?
> Eh?
>
> or are protocols in IP not routeable?
> Eh?
>
> if both are, why do we have to consider BGP into OSPF distribution?
> Depends on the design and requirements of the network; ISPs generally
> don't distribute BGP routes into the IGP whereas an enterprise using
> BGP as the global core might (again, depending on the network design)
> gain more accurate routing from doing so.
>
> why not L3VPN running OSPF or RIP?
> Where? PE-CE? You can -
>
> very *grey* question, but would appreciate if someone could clarify.
>
> -rgds
> Alok
>
>
> -------
> The MPLS-OPS Mailing List
> Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml
> Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
>
> -------
> The MPLS-OPS Mailing List
> Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml
> Archive: http://www.mplsrc.com/mpls-ops_archive.shtml



-------
The MPLS-OPS Mailing List
Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml
Archive: http://www.mplsrc.com/mpls-ops_archive.shtml



Yahoo! Plus - For a better Internet experience