Ping between routers

[Pietro955]

Hi guys,

I have a doubt, and I'd like if anyone of you for sure more expert than me can show all the packets life in pingig R 3 from R 1. There are no static routes, and the interfaces are all up. I'll be really grateful if you can show me the entire process for processing the ping request from R 1 to R 3, including all the ARP packets. I cannot figure out how it can ping.

thank you very much

regards

[Pietro955]

Sorry, I just saw that there are 2 static routes: one on R1 for R3, and one on R3 for the route that can reach R1. So my thoughts are:

when I ping from R1, R1 knows that has to send that packet to the interface that can reach R3. R1 tho doesnt know the MAC of that interface so he keeps the ICMP packet and sends an ARP request. That interface answer with an ARP reply and its MAC address, then the router sends the packet to that interface. Tell im please if I'm right or not... Thank you

[Giuseppe Larosa · Accepted solution]

Hello @Pietro955 ,

>> Sorry, I just saw that there are 2 static routes: one on R1 for R3, and one on R3 for the route that can reach R1

in this case R1 will send out an ARP request for R2 IP address on 10.10.10.0/24 in order to be able to send the frame to R2.

R3 will do the same for R2's IP address on 10.10.20.0/24 subnet.

Hope to help

Giuseppe

[balaji.bandi]

your question how you can achieve routing reachability between R1 and R3 without any static Route?

Thinking that all your switches are Layer 2

Then you need to choose any of the IGP protocols,  which can dynamically learn the route from a neighbor peer router.

1. R1 learn routes from R2

2. R2 Learn route from R1

3. same way R3 and R2

in your case R2 is transit, so R1 try to  reach R3 IP address, R1 next hop is R2, R2  will get that request and look in the routing table and send egress where this route belongs towards R3.

This high level - you can find many document deep dive of routing learning.

[Giuseppe Larosa]

Hello @Pietro955 ,

>> There are no static routes, and the interfaces are all up

ok assuming there is no routing protocol in use and no static default route on R1 pointing to R2 IP address the only way to make this work is to rely on proxy ARP.

R1 needs to have a much larger subnet mask like 255.255.0.0 associated to the LAN interface.

At this point when the ICMP echo request SA= 10.10.10.1 DA = 10.10.20.1 is isued on R1 console the following happens:

R1 thinks that 10.10.20.1 is in the same subnet and will send out an ARP request òlooking for the MAC address of 10.10.20.1.

R2 receives this broadcast frame. if R2 has proxy ARP enabled it will answer to the ARP request with an ARP reply that contains the MAC addres of its own interface fas0/0.

At this point R1 sends the ICMP packet to R2 using the appropriate MAC addresses

R2 processes the ICMP packet and sends out and ARP request for destination 10.10.20.1.

The same things happen on the lan segment between R2 and R3

R3 needs to have a larger subnet mask like 255.255.0.0

R2 needs to have proxy ARP enabled on its other LAN interface.

Proxy ARP was used for helping hosts with broader subnet masks to be able to talk with hosts in other subnets.

R2 in this casae helps the devices to communicate.

Note that if R1 subnet mask is not larger communication would not be possible as R1 would not ARP for destination address ! 

Another possible option is to disable ip routing on R1 and R3.

no ip routing

ip default-gateway 10.10.10.2

So either you rely on proxy ARP or you need to disable routing and to use the ip default-gateway concept.

Hope to help

Giuseppe