A few weeks ago, Urs Baumann posted a nice example illustrating the power of netlab: a 10-router topology running OSPF, IS-IS, and BGP:

He didn’t post the underlying topology file, so let’s create a simple topology to build something similar.

defaults.device: frr
provider: clab

• We’ll use FRRouting-based routers (line 1) because we want to be able to run this thing in GitHub Codespaces.
• We’ll have to run them in containers (line 2).

groups:
  _auto_create: True
  c_65000:
    module: [ bgp ]
    members: [ r01 ]
    bgp.as: 65000
  p_65100:
    module: [ ospf, bgp ]
    members: [ r02, r03, r04, r05 ]
    bgp.as: 65100
  p_65200:
    module: [ isis, bgp ]
    members: [ r06, r07, r08, r09 ]
    bgp.as: 65200
  c_65002:
    module: [ bgp ]
    members: [ r10 ]
    bgp.as: 65002

The easiest way to create a large number of similar nodes is to use node groups:

• Group members will be created automatically (line 5)
• The first group (left customer) has a single member running BGP in BGP AS 65000 (line 6-9).
• The second group defines the left provider (lines 10-13). It has four members running OSPF and BGP in AS 651000.
• The rest of the above defines the right provider (lines 14-17) and the right customer (18-21)

We also need links between routers. Yet again, we’ll use link groups to create them:

links:
- group: p_65100
  prefix.ipv4: True
  members: [ 
    r02-r03, r02-r03, r02-r04, r02-r04, 
    r03-r05, r03-r05, r04-r05, r04-r05 ]
- group: p_65200
  prefix.ipv4: True
  members: [
    r06-r07, r06-r08, r07-r09, r08-r09 ]
- group: interas
  members: [
    r01-r02, r01-r03,
    r04-r06, r05-r07, 
    r08-r10, r09-r10 ]

• The first group defines unnumbered IPv4 links within AS 65100 (lines 24-28). The member list is pretty long, as Urs wanted to have two parallel links between pairs of routers (and I love how one could write lists in YAML ;)
• The second group defines unnumbered IPv4 links within AS 65200 (lines 29-32). This group has only four members.
• The final group (lines 33-37) defines all links between autonomous systems.

That’s it. Save the file as topology.yml and execute netlab up. Don’t want to invest time into building your own netlab infrastructure? Use GitHub Codespaces:

• Start a new GitHub Codespace
• Change directory to routing/multi-as
• Execute netlab up and wait a bit
• Wait for another 30 seconds to get the routing protocols up and running.
• Connect to r01 with netlab connect r01 and execute a traceroute:

r01(bash)#traceroute -s 10.0.0.1 r10
traceroute to r10 (10.0.0.10) from 10.0.0.1, 30 hops max, 46 byte packets
 1  eth5.r02 (10.1.0.2)  0.003 ms  0.001 ms  0.001 ms
 2  r04 (10.0.0.4)  0.001 ms  0.001 ms  0.001 ms
 3  eth3.r06 (10.1.0.10)  0.001 ms  0.001 ms  0.001 ms
 4  r08 (10.0.0.8)  0.000 ms  0.000 ms  0.001 ms
 5  r10 (10.0.0.10)  0.001 ms  0.000 ms  0.001 ms

Not only did netlab create a running network for you, but it also set up reverse DNS mappings to display the (numbered) interfaces your traceroute probes are traversing.