  yaplej CCNA Premium join:2001-02-10 White City, OR
·Charter Pipeline
·Clearwire Wireless
| [Config] BGP and OSPF redistribution
I'm going through my first BGP configuration that I had to configure for our new MPLS WAN.
I have some questions about whether I should redistribute our OSPF, and static routes into BGP though. Iv read some stuff that says I should not, but it seem to be mostly related to using BGP for advertising routes to the Internet not for internal routing.
I have been successful at advertising a default route into the MPLS network, but the routes from our frame, and ppp network are not being advertised to the MPLS network through BGP although the routes do exist on the router connecting to the MPLS network.
Thanks for you thoughts. |
|
  dpocoroba Premium join:2000-11-14 224.0.0.5
| First thing that comes to mind is "KIS" Keep it simple. You need to address some thing to make that decision. Does you network route and function correctly now? if yes then why change it.
If that answer is no, then what exactly are you trying to accomplish? Even with only a few routes being advertized you should be able to achieve any type of redundancy, etc.
IP schemes also come into play here. If everything is done properly then you should have the least amount of routes as possible. There is nothing wrong with advertising only a handful of routes on a MPLS network. You could also look at summary routes from multiple /24's to a single /16 or even just a default route.
It will all come down to what you want and are trying to accomplish. I am just offering my real world advise from doing several MPLS networks and what I know works.
All my sites advertise just 1 route with the exception of the 2x data centers which advertise just 2x routes each to the MPLS.
DP -- "Knowledge is contagious, infect" |
|
  yaplej CCNA Premium join:2001-02-10 White City, OR
·Charter Pipeline
·Clearwire Wireless
| reply to yaplej I know that the MPLS network is receiving the default route so that's good. None of the other routes from our point-to-point or frame network are in the MPLS network though.
I'm a little worried that after I move these locations over to the MPLS network then the other sites still on the point-to-point and frame relay wont know how to route to the networks on the MPLS network.
I was hoping to get routes for networks in the MPLS network to advertise through the ppp/frame, and also the routes for networks on our ppp/frame to be advertised into the MPLS network.
I guess Ill find out when I actually get another network connected to the MPLS network. I'm just trying to lean as much about it as possible to make the transition easier. I can always add static routes if it doesn't work. |
|
  rolande Certifiable Premium,Mod join:2002-05-24 Powell, OH clubs:
Host: Linksys AT&T Midwest
| Static routing is always bad the larger the network gets.
Are you moving your entire branch network into the MPLS cloud? Depending on the rest of your network architecture, you really only need to send a default or summaries to the MPLS BGP process. If there is only one path to and from your network into the MPLS environment then I wouldn't worry about it.
Now if you are trying to do tricky things with multiple MPLS pipes and split load between 2 different links, then that warrants further analysis as to the proper way to address the routing to meet those requirements. -- Ignorance is temporary...stupidity lasts forever!
»www.thewaystation.com/ »blog.thewaystation.com/ |
|
  sporkme drop the crantini and move it, sister Premium,MVM join:2000-07-01 Budd Lake, NJ | reply to yaplej It's probably time for a picture. |
|
  yaplej CCNA Premium join:2001-02-10 White City, OR
·Charter Pipeline
·Clearwire Wireless
| reply to yaplej We are moving the entire network to MPLS we just got started with three locations though.
I got everything working yesterday, and the only real issue I had is that our firewall was not receiving the routes to networks located in the MPLS cloud. So I put in a static route in the firewall until we figure out how to fix that.
I could setup BGP between the 4305, and the firewall, but I'm not sure if that's how I should do it, or if I should just redistribute between OSPF & BGP, or something different all together.
Thanks. |
|
  rolande Certifiable Premium,Mod join:2002-05-24 Powell, OH clubs:
Host: Linksys AT&T Midwest
| If you are not summarizing into OSPF for your MPLS cloud / BGP subnets, then you have to get the specific routes there somehow or else they stay in BGP and none of the rest of your network learns about them. The statics will work on the firewall as you transition but it probably isn't the best solution.
Are you summarizing from the existing frame cloud today? I would remove the summary and allow all the subnets to advertise until the migration is complete. If you are not already summarizing, which I suspect you are not, then I would advertise the summary. As you move each branch into the MPLS cloud, the more specific route will disappear from OSPF and the summary will become the preferred path which will then hit your router that is running the MPLS BGP process and has the more specific routes in its table. -- Ignorance is temporary...stupidity lasts forever!
»www.thewaystation.com/ »blog.thewaystation.com/ |
|
 aryoba Premium,MVM join:2002-08-22 | reply to yaplej yaplej,
You probably need to post the running configurations of the PE, CE, and non-MPLS routers (that still in non-MPLS network). This way we should have all the info necessary. |
|
  rolande Certifiable Premium,Mod join:2002-05-24 Powell, OH clubs:
Host: Linksys AT&T Midwest
| aryoba , more than likely he does not even manage the CE MPLS router, let alone the PE router. The provider manages everything and he only manages a router that performs ethernet handoff with a BGP peering session. That is typical deployment unless they are running native MPLS within their network, which I doubt.
Configs for the 4305 would be helpful though. |
|
  yaplej CCNA Premium join:2001-02-10 White City, OR
·Charter Pipeline
·Clearwire Wireless
| reply to yaplej Just an update.
I was able to redistribute our OSPF routes into our MPLS cloud finally. I had to upgrade to a newer version of firmware on the Adtran, but after that everything is working great.
The MPLS cloud can see the default route originating from our ASA5500, and all the routes still on the frame/ppp network also. |
|