Couple of BGP questions...

Unanswered Question
Sep 10th, 2008
User Badges:

1. I know the standard L0 address is a /32, but can it work if you use any address out of a /28 for example? Any benefits or drawbacks to doing it either way?


2. I'm converting my org's internet connections from static to using BGP since we'll now be multi-homing to two different providers. One connection with our new provider is up and running as expected. On the other connection, could I potentially configure the BGP config and leave the static in place until we're ready to cut over with our ISP? My thoughts are that the static would have preference over the protocol and not cause any issues. The other BGP speaker is not currently taking any traffic since we're advertising a /24 we're not yet using.


Thanks in advance.


/rls

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Giuseppe Larosa Wed, 09/10/2008 - 12:02
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

Hello Robert


1) you can use a /28 but all those addresses will be wasted. Loopback with /2x are used as placeholder to advertise NAT pools for example and pools of ip addresses assigned to ppp remote users by NAS devices.


2) be aware to avoid that the static routes having a natural AD of 1 <20 of eBGP could take precedence also on the current active eBGP neighbor I would use floating static routes with AD > 20 so that when eBGP sessions will be used they will take precedence.


Hope to help

Giuseppe


Richard Burts Wed, 09/10/2008 - 12:13
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

rls


1) I am not clear whether you are suggesting that the loopback would be assigned a /28 or whether you are suggesting using an address out of a /28 that is not the loopback. Can you clarify.


Additionally I would comment that the use of loopback address for BGP peering is especially common in IBGP. The common practice for EBGP is to peer from the physical interface address on which the BGP peer connects.


2) I think we can not give you very good answers without knowing a bit more about your network. Are you currently using just static routes for routing within your network or are you using some dynamic routing protocol? How many routers are there in your network?


A static route will take precedence over a BGP route for the same prefix. But an EBGP route would take precedence over a route from the common Interior routing protocols like EIGRP or OSPF.


Another aspect of the issue is what static routes you currently have (I suspect that it is just a static default route) and what routes the provider will begin to advertise to you. If they start advertising some prefixes from their network to you, then their advertisement will be a more specific prefix than your default route. And their more specific prefix will take precedence over your static default route.


Another thought is that your decision about whether to run BGP with the existing provider may depend on how they are willing to do things. It is highly likely that now they have a static route for your network. When they enable BGP for your network they may want to remove their static route.


HTH


Rick

Actions

This Discussion