point to point and inverse arp?

mguy

Why if i have two routers connected to a frame relay, they do not learn about each other?

But if I have three routers in a point-to-multipoint network, then they do learn each other dynamically?

elderkai

I'm not really sure what you're asking. Learn about each other how?

mguy

learn about each other through inverse arp.

On the two router scenario, they have each other's interface configured dynamically through inverse arp, but this is not the case when there are three on a hub-spoke topology.

elderkai

Inverse ARP just isn't practical and perfect. There are a lot of problems like that happen to it, so I would suggest disabling inverse arp right after enabling frame-relay encapsulation and doing everything statically.

Zartanasaurus

mguy wrote: »

learn about each other through inverse arp.

On the two router scenario, they have each other's interface configured dynamically through inverse arp, but this is not the case when there are three on a hub-spoke topology.

Spokes aren't going to learn about each other dynamically because InARP messages aren't forwarded by the hub. The hub will learn all of the spokes dynamically. The spokes need static mappings. Or you can use P2P subinterfaces on the spokes which doesn't use InARP.

mguy

Zartanasaurus wrote: »

Spokes aren't going to learn about each other dynamically because InARP messages aren't forwarded by the hub. The hub will learn all of the spokes dynamically. The spokes need static mappings. Or you can use P2P subinterfaces on the spokes which doesn't use InARP.

Thanks. So if I use a hub-spoke topology, then inverse arp is activated and the remote interfaces are learned dynamically.

As opposed to in point-to-point, the hub interface have to mapped manually, because inverse map is not used.

Question:

how does the hub interface know that I have multiple spokes as opposed to a single spoke (which is a point-to-point)?

If it were a point-to-point connection through a frame-relay, woudn't a inverse-arp mechanism just be received by the single remote router and configure dynamically--- why the difference?

Yeah, so how does it know?

NetworkVeteran

Or you can use P2P subinterfaces on the spokes which doesn't use InARP.

mguy wrote:

how does the hub interface know that I have multiple spokes as opposed to a single spoke (which is a point-to-point)?

He was talking about point-to-point subinterfaces. A router knows if you type--

int s0/0.100 point-to-point

mguy wrote:

why the difference?

A point-to-point subinterface has only one DLCI, so the IP-to-DLCI mapping is obvious.

mguy

Actually I tested this theory on my packet tracer, and the routers configured themselves dynamically in BOTH point-to-point and in point-to-multipoint.

?????????????

mguy

No I didn't use subinterfaces in my tests.

NetworkVeteran

mguy wrote: »

No I didn't use subinterfaces in my tests.

Then you didn't test point-to-point Frame Relay subinterfaces, which was what he was talking about.

Here's a helpful webpage from Cisco--

http://www.cisco.com/en/US/tech/tk713/tk237/technologies_tech_note09186a008014f8a7.shtml#topic4

mguy

recap.

interface are default multipoint

if subinterface multipoint, inverse is used so dynamic mappings
if subinterface point-to-point, inverse is not sent so static mappic.

now time for routing protocols and their behaviors under these settings

mguy

Spokes aren't going to learn about each other dynamically because InARP messages aren't forwarded by the hub. The hub will learn all of the spokes dynamically. The spokes need static mappings (of other spokes). Or you can use P2P subinterfaces on the spokes which doesn't use InARP.

I think i got it. Thank you!!