Skip to Content.
Sympa Menu

wg-multicast - Re: how to make this work?

Subject: All things related to multicast

List archive

Re: how to make this work?


Chronological Thread 
  • From: Marshall Eubanks <>
  • To: David Farmer <>
  • Cc: wg-multicast List <>, Charley Kline <>, debbie Fligor <>
  • Subject: Re: how to make this work?
  • Date: Sat, 2 Feb 2008 13:27:52 -0500

Dear David;

On Feb 1, 2008, at 8:58 PM, David Farmer wrote:

Marshall,

Thanks for the pointer, can you provide more info on the two types?

Are you aware of implementations other that Cisco and Juniper?

I believe that Alcatel has one (this came out at the last IETF).

I did not point you to the latest version, for which I apologize. Here it is :

http://tools.ietf.org/id/draft-ycai-mboned-mvpn-pim-deploy-00.txt


Personally, it would be really nice if some of this actually made it from Drafts
to even just Informational RFCs.

Others,

If you are interested in mVPN you should read this draft and many of the
references in it.

FYI, for the record, we implemented our mVPNs with PIM-SSM for the P

PIM-SSM or PIM-SM ?

Regards
Marshall

network, AKA the default-MDT and data-MDT or MI-PSMI and S-PMSI. Our
Border Routers are the RP for the "customer" PIM-SM and connect the
mVPNs to each other and the Internet.

If RFCs are not your style, maybe try this;

http://searchnetworking.techtarget.com/searchNetworking/Downloads/MPLS
ch7.pdf


On 1 Feb 2008 Marshall Eubanks wrote:

Have you talked to your vendor ? Are you aware of the two flavors of
MPLS multicast ?

This draft might be useful if you are doing PIM MPLS Multicast :

http://merlot.tools.ietf.org/html/draft-ycai-mboned-mvpn-deploy-00

If it _is_ useful to you (or, even if it isn't), please let me know,
as we still haven't decided whether or not
to make an official draft or not.

Regards
Marshall



On Feb 1, 2008, at 1:00 PM, debbie fligor wrote:

We're doing some planning for our next-generation campus core
network. We're planning on using VRFs to provide some specialty
services (like a high-speed research network that bypasses the exit
firewalls).

Here's a very simplified version of what we're thinking it will
look like:

<mcast-core6.jpg>


and an ascii version for those that don't do inline jpegs:

vrf blue net 3 --- (dist router) -- (core router) -- (dist router)
--- vrf blue net 4
/ /
vrf red net 1 ----/ vrf red net 2/


We want to be sure that Intra-VRF multicast (eg. from net 3 to net
4, all in the "blue" vrf) will work if the links between the dist
routers and core routers are MPLS based. We're confident we can
make it work if the links between the core and dist routers are L2
and we use VRF-lite, but we are interested in considering MPLS
instead.

We're assuming we'll be able to make the Inter VRF multicast work,
even if it's a the other side of our exit, by treating them as L3
peerings.


So if you've done something like this, how did you do it? So far
it looks like we'd have to run BGP as a core routing protocol and
exchange NLRIs with MBGP, if that's even defined over MPLS (I don't
think the standard for that is done, info on that is also
appreciated). We're not sure BGP makes an ideal core routing
protocol for a campus.

We're especially interested in hearing how you have made something
like this work (or not!) on Cisco 6500's and Juniper MX-480 or
MX-960. Anything you've done with other vendor's gear that you
think might be insightful is also appreciated.

thanks!


-----
-debbie
Debbie Fligor, n9dn Network Engineer, CITES, Univ. of Il
email:

<http://www.uiuc.edu/ph/www/fligor>
"My turn." -River Tam




=================================================
David Farmer Email:

Office of Information Technology
University of Minnesota Phone: 612-626-0815
2218 University Ave SE Cell: 612-812-9952
Minneapolis, MN 55414-3029 FAX: 612-626-1818
=================================================





Archive powered by MHonArc 2.6.16.

Top of Page