Basic Default Route ?

[JoshGarmonCiscoCCNA]

Interior------------------fe 0/1  RouterA ser 0/0/0------------------------ser 0/0/1 RouterB---------------------*Internet*

Where do I configure the default route again so that Router A knows what's coming from and should go to the Internet?

[Peter Paluch · Accepted solution]

Josh,

I am not entirely certain I understand your question, but in general:

RouterA would point its 0.0.0.0/0 out its Serial0/0/0 toward RouterB. RouterB would point its 0.0.0.0/0 toward its upstream router in the internet.

If you used a dynamic routing protocol, it would be sufficient to configure the 0.0.0.0/0 only on RouterB, and have this default route advertised from RouterB to RouterA via this routing protocol. RouterA would learn the default route just like any other network, it would properly choose RouterB as the next hop toward it (as that's the router that told RouterA about it in the first place), and would propagate it to further routers if there were any.

Best regards,
Peter

[Joseph W. Doherty]

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

"RouterA would point its 0.0.0.0/0 out its Serial0/0/0 toward RouterB."

Just to add a footnote to what Peter has described, I'm sure Peter is talking "logically". "Physically" a static route statement could just point to the interface, alone, but it's generally much better to use the next hop address, with, or without, the interface.

[Peter Paluch]

Joseph,

it's generally much better to use the next hop address, with, or without, the interface.

I absolutely agree and I am myself an avid proponent of configuring static routes with next hop IP addresses instead of egress interfaces whenever possible. However, in this particular case, I didn't want to go too deep into these intricacies, and I also assumed that because Josh's topology mentioned the use of Serial interfaces that by default run HDLC, a point-to-point protocol in Cisco's implementation, it is safe to refer to the egress interface for the sake of simplicity.

Good point!

Best regards,
Peter

[Joseph W. Doherty]

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

Peter, as your typical, a wonderful response.

"However, in this particular case, I didn't want to go too deep into these intricacies, . . . "

Indeed, there's always the question of too little or too much information in a response.

". . . and I also assumed that because Josh's topology mentioned the use of Serial interfaces that by default run HDLC, a point-to-point protocol in Cisco's implementation, it is safe to refer to the egress interface for the sake of simplicity."

Well, I did write "generally" not always - laugh

That said, so default routing to a serial p2p, using a default, doesn't ARP?  I didn't realize that.  Of course, I can see why it shouldn't need to.  Same for a /30 or /31 on an Ethernet port?  I.e. is the key the media or the network?

[Peter Paluch]

Hi Joe,

That said, so default routing to a serial p2p, using a default, doesn't ARP?

Quite so, it doesn't. This is actually a part of a broader context with point-to-point Layer2 technologies. On a point-to-point technology, there is no reason to do any kind of L3-to-L2 mapping because there is always exactly one peer reachable over a point-to-point interface, never two or more, and the corresponding Layer2 technologies take this into account by not using any specific distinct Layer2 addresses. If you dissect the PPP or Cisco HDLC frame, while there is an Address field, it is always set to a constant value (Cisco HDLC uses 0x0F, PPP uses 0xFF). With Frame Relay on a point-to-point subinterface, there is a sensible Layer2 address - the DLCI - but the DLCI is bound to the subinterface by its configuration and never changes dynamically, so even here, there is nothing to do in L3-to-L2 mapping, as the resulting Layer2 address is constant at all times.

This actually goes so far that for every communication over a point-to-point interface, the following paradigm is used: "Every address from the network directly attached to that interface is on the other side." If you ping your own address on a point-to-point interface, the packets will actually be sent out the interface, received by the other side and the other side - using the same logic - will send them back to you. So if you ping yourself, the packets actually go out the point-to-point interface, up to the neighboring device, and back to you.

is the key the media or the network?

The key is the nature of the underlying Layer2 technology of the interface. Every point-to-point Layer2 technology (Cisco HDLC, PPP, Frame Relay using point-to-point subinterfaces, ATM using point-to-point subinterfaces, X.25 using point-to-point subinterfaces) will operate this way, regardless of what the IP netmask on the interface is. On the contrary, if you take Ethernet, it is by nature a multi-access technology, so regardless of the mask, it will do ARP because it is the only sensible way to populate the Destnation MAC address field.

It could be said, with a degree of simplification, that you don't need ARP (or any similar mechanism for that purpose) if the underlying Layer2 technology uses a constant value to address the (only) peer.

I guess I got this far more complicated than I originally intended... sorry for that.

Best regards,
Peter

[Joseph W. Doherty]

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

"I guess I got this far more complicated than I originally intended... sorry for that."

Yup, longer than I expected, but excellent - thanks again - others should appreciate too.

[JoshGarmonCiscoCCNA]

Is that where the command

default-information originate comes into play?

[Peter Paluch]

Hi Josh,

Yes, exactly. You would use the default-information originate command in RIP, OSPF or IS-IS to make your particular router inject a default route into the list of networks it is advertising. Other routers would not be configured with this command as they would not be injecting their own default route; rather, they would only be learning about it and propagating it further.

Notably, EIGRP does not support the default-information originate command (there's no logical explanation for that, Cisco simply didn't consider it to be important enough). Instead, you would use the redistribute command to retake the default route from its original source (say, the static configuration) and import it into EIGRP. After that, it would be the same as with the remaining protocols.

Best regards,
Peter

[Jon Marshall]

Peter

Just for your info, because I'm guessing you would want to know this, Cisco have introduced the default-information originate command for EIGRP on Nexus switches.

Perhaps they will be adding it to the other OS versions at a later date.

Jon

[Peter Paluch]

Jon,

Just for your info, because I'm guessing you would want to know this, Cisco have introduced the default-information originate command for EIGRP on Nexus switches.

Did they??? I AM CRACKING THE CHAMPAGNE!!! :-P

It really was about time. Just think of it - the code base with all other routing protocols is already there. One wonders why someone didn't implement it in IOS simply out of the feeling of asymmetry in the supported features... :)

Thanks for letting me know!

Best regards,
Peter