Added FRR "Transit AS" Labs

[DecioXXIV]

I tried to realize the FRR version for the "Transit AS with BGP Redistribution in RIP" Lab.

Unluckily, this lab behaves in a non-deterministic way, so my solution is still partial.

[lorenzo93]

Hi @DecioXXIV ,

thanks very much for your work, we appreciate it a lot.
Please, correct me if I'm wrong, but according to the folder name I understand that you transpose to FRR the "non-deterministic" version of the lab so, the non-determinist behaviour is expected.
Moreover, there are some differences from the original quagga lab that I'll highlight in the review. I've made some comments just on one file but they should be seen as "global" comments.

In addition, there are no default routes in the original lab, while yours has many of them. Please remove it.

[DecioXXIV]

Thanks for your comments. I will fix everything you suggested me.

Moreover, the non-deterministic behavior is mainly related to as10rt2 and as10r5.
To realize the Transit AS, it is required that as10rt2 use as10rt3 as the egress point from the AS, implying that packets to 100.0.0.0/8 will be routed to as10rt3.

At the same time, as10rt2 redistributes BGP in RIP. The non-determinism is related to which is the Next-Hop to 100.0.0.0/8 in the Routing Table of as10r5.
Starting the Lab several times, sometimes it happens that as10r5 chooses 11.0.0.6 (as10rt2) as the Next-Hop, in the other cases it chooses 10.0.0.6 (as10r6).
In this situation, when as10r5 routes packets to as10r6 nothing bad happens, but when it routes packets to as10rt2 instead, these are bounced between as10r5 and as10rt2 themselves.

[DecioXXIV]

Hi, I updated the Lab files with the corrections you suggested me.

I'm still working to find a fully-deterministic solution.
I should be on a good way, hope that my next commit will be providing a potentially definitive solution.

[DecioXXIV]

Hi, I pushed the deterministic version of the Lab. It should be correctly working.

Just a little insight: as you can see in the "frr.conf" file of Router as10rt2, I decided to block the redistribution in RIP of the prefix 100.0.0.0/8 (even if learned by eBGP).
This because its redistribution could again cause the bouncing of packets directed to 100.0.0.0/8 between the Routers as10rt2 and as10r5.

[DecioXXIV]

Hi, I pushed the version of the "Transit AS" Lab which exploits the Tunneling. It should be correctly working.