Routing/redistributing to a network on the same subnet.

Answered Question
Mar 19th, 2009
User Badges:

Check the attachment to see if this setup looks valid.


Senario:

You have a connection to a business partner. You're router connects to the business partners router, both your routers interface and the business partners router interface are on the same subnet as should be. This same subnet/network the interfaces are part of is the subnet you are routing to on the business partners side. Does this look like a valid configuration. You are going to be redistributing the static route into you table (EIGRP), the vendor will be just using static routes. My questioning this is the way I have usually set this up is to use like a /30 on the p-to-p interfaces with a separate network routing to that is not in the same subnet as the interfaces.


Thanks,

Brandon



Correct Answer by Jon Marshall about 8 years 3 months ago

Brandon


"Thanks so as long as the person who has control of router R2 sets the static and as long as the R1 side has 192.168.255.0 0.0.0.127 (where the fa0/0 is 192.168.255.129/25) under it's EIGRP process traffic would arp out?"


Kind of. R1 arping out has nothing to do with whether there is a 192.168.255.128 (note .128 not .0) entry under the EIGRP process. R1 will arp out because 192.168.255.128 is locally connected.


The entry under the EIGRP process is only needed if there are other routers behind R1 that need to know about the 192.168.255.128 network. If there aren;t any routers then you don't need to run EIGRP on R1 at all.


Bear in mind also that in addition to the static route on R2 you also need all the NAT configuration for the hosts on the 121.0.0.0/8 network.


Jon

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Jon Marshall Thu, 03/19/2009 - 17:53
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Brandon


If i understand your question correctly no this isn't a valid configuration as each router has the P2P link configured as a /25 which is also the business partner network.


So each router considers the 192.168.255.128/25 network to be directly connected which will take precedence over the static route configured.


Use a separate network for P2P link as you usually do.


Jon

Joseph W. Doherty Thu, 03/19/2009 - 18:02
User Badges:
  • Super Bronze, 10000 points or more

As an alternative, you might be able to bridge the 192.168.255.128/25 network across the partner's router. (This assume the partner router doesn't need to route elsewhere.)


If you do bridge, your business router would be the gateway for 192.168.255.128/25 network and you could just define 192.168.255.128/25 to EIGRP (you wouldn't need a static or redistribution.)


(NB: If the partner router is Ethernet both sides, and it doesn't have to otherwise route, you could also use a L2 switch.)


PS:

Jon, recall your suggestion of router bridging way back - if this doesn't seem right, please correct.

Jon Marshall Thu, 03/19/2009 - 18:14
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Joseph


We may be reading this slightly differently. I don't think Brandon needs to have the 192.168.255.128/25 network on both sides. If he did then yes bridging would be a solution.


But the business partner network is 192.168.255.128/25 and the business network is 10.101.0.0/16. The problem is that the P2P link between the 2 routers is configured out of the 192.168.255.128/25 network.


If i have understood Brandon correctly then the P2P network just needs changing to fall outside of the 2 business networks.


Jon



Joseph W. Doherty Thu, 03/19/2009 - 18:51
User Badges:
  • Super Bronze, 10000 points or more

Jon, agree the existing problem is the link between the two routers, i.e. same /25 on both sides of the partner router. Also fully agree using another network between the two routers is a solution.


I futher agree Brandon may not need to have the 192.168.255.128/25 network on both sides, but does it work if you do by bridging the network?


As you also note, the two networks are, partner, 192.168.255.128/25 and, business, 10.101.0.0/16. But given just that, there's really no need(?) for the partner router if we keep 192.168.255.128/25 connected to the business router. Futher, if we include the business router's 192.168.255.129 interface with EIGRP, we eliminate the need for a static route and redistribution.


I'm not suggesting this is a better solution, although it saves on a couple of IP addresses (router interfaces) and a new network (between routers) allocation (and also may allow replacement of the partner router with a L2 switch).


Again though, is this a possible solution? If the business partner router were to be retained, from what I'm describing, the partner router would need to bridge its two interfaces, correct?

Jon Marshall Fri, 03/20/2009 - 12:30
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Joseph


Apologies for delay in getting back.


If you removed the partner router then yes you wouldn't need the static route but i would think it unlikely any partner would be happy to connect straight into their network. I suppose they could use a firewall to terminate the connection but the firewall would have to be in transparent mode otherwise you are back to the L3 next-hop issue again.


If the partner router stayed then yes you could bridge the 2 interfaces together although it does depend on the speed of the link ie. slow links don't really bridge that well.


Overall i would only look to bridge if the addressing really couldn't be changed and as it is 192.168.x.x addressing i'm assuming it is.


NAT could also be a possible solution ie. present the remote 192.168.x.x addresses as something else on the partner router.


Out of all of those i'd definitely go with readdressing P2P :-).


Jon

mbroberson1 Mon, 03/23/2009 - 06:44
User Badges:

Hi Jon,


What if you configured it so your router and the adjacent router both have an address from the same subnet. This would also be from the same subnet behid the adjacent router, but on your router you would put the network number under the EIGRP process. You think this would work?


Thanks

Jon Marshall Mon, 03/23/2009 - 08:38
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Brandon


Not 100% sure i understand. If you mean


R1 (192.168.5.1/24) -> (192.168.5.2/24) R2 (192.168.5.3/24)


then you couldn't do this as R2 would complain about overlapping subnet when you tried to configured the 2nd interface with an IP out if the same 192.168.5.x network.


You could bridge the 2 interfaces on R2 which is kind of what Joseph was suggesting and then give a 192.168.5.x address to the bridged interface.


This is however just making things more complicated than they need to be unless of course you cannot change the existing IP addressing.


Jon

mbroberson1 Mon, 03/23/2009 - 08:49
User Badges:

Hi Jon,


I mean more like this:


R1 - fa0/0 - 192.168.255.29/25

connected to:

R2 - fa0/0 - 192.168.255.30/25

Host are addressed (as part of another network (that of the R2 fa0/1 interface) and being nated to a 192.168.255.128/25 network address) behind the R2 fa0/1 interface and being nated to R1.


Under the R1 EIGRP process you would specify 192.168.255.0 0.0.0.127


R2 is using static routing.


Thanks,

Brandon

Jon Marshall Mon, 03/23/2009 - 12:45
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Brandon


Sorry you have kind of lost me on this. Could you do a quick diagram of what you are trying to achieve.


The bit where you lost me was "behind the R2 fa0/1 interface and being nated to R1"


Jon

Jon Marshall Mon, 03/23/2009 - 13:40
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Brandon


That makes sense now.


Yes this would work. As you say you would need to add a route for the 10.10.10.x/24 network to the router on the right in your diagram (R2).


R1 (on the left) would simply arp out for any 192.168.255.128/25 address it needed.


Jon

mbroberson1 Mon, 03/23/2009 - 13:43
User Badges:

Jon,


Thanks so as long as the person who has control of router R2 sets the static and as long as the R1 side has 192.168.255.0 0.0.0.127 (where the fa0/0 is 192.168.255.129/25) under it's EIGRP process traffic would arp out?


Sometimes my diargams are more easily explained than words...;-)


Thanks,

Brandon

Correct Answer
Jon Marshall Mon, 03/23/2009 - 13:50
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Brandon


"Thanks so as long as the person who has control of router R2 sets the static and as long as the R1 side has 192.168.255.0 0.0.0.127 (where the fa0/0 is 192.168.255.129/25) under it's EIGRP process traffic would arp out?"


Kind of. R1 arping out has nothing to do with whether there is a 192.168.255.128 (note .128 not .0) entry under the EIGRP process. R1 will arp out because 192.168.255.128 is locally connected.


The entry under the EIGRP process is only needed if there are other routers behind R1 that need to know about the 192.168.255.128 network. If there aren;t any routers then you don't need to run EIGRP on R1 at all.


Bear in mind also that in addition to the static route on R2 you also need all the NAT configuration for the hosts on the 121.0.0.0/8 network.


Jon

mbroberson1 Mon, 03/23/2009 - 17:49
User Badges:

Hi Jon,


Man it seems like you know this setup! As for your 3rd paragraph, yes there are other routers behind R1 (running EIGRP) that need to know about the 192.168.255.128/25. Also the vendor in-charge of the R2 router will be doing their static routing in-addition to the nats.


I am going to post this as a working solution.


Kind Regards,


Brandon

mbroberson1 Fri, 03/20/2009 - 11:44
User Badges:

Hi Jon,


You're are correctly understanding the question/senario and you have answered it the way I thought it should be. I just needed clarification on the setup (in case I was missing something my self). A vendor I am working with possed this solution and I thought it looked invalid, but wanted the pros to voice opinion.


Thanks!


Brandon

Actions

This Discussion