The scenario
ISP-A (AS 65001), ISP-B (AS 65002) and ISP-C (AS 65003) each have one port on IXP-Switch and an address on the exchange network 10.100.0.0/24. Each ISP runs a separate eBGP session with each of the other two across that switch, and each has one customer: Customer-A on 172.16.1.0/24, Customer-B on 172.16.2.0/24 and Customer-C on 172.16.3.0/24. No router has a static route and none runs OSPF, so every route to another ISP's customer is one that ISP advertised. The switch only carries Ethernet frames between the three router ports; it has no address and makes no routing decisions.
What you will work out
Understand how ISPs peer across a shared exchange switch, why the switch does no routing, and how BGP picks between a direct path and a longer one.
Why this matters
Networks that send each other a lot of traffic often meet at an exchange instead of paying a transit provider to carry it, which saves money and usually shortens the path. What happens at the exchange is exactly this lab: a shared switch, one BGP session per peer (or a route server), and routes that exist only while those sessions are up. When a peer's port goes down, its routes are withdrawn and traffic takes whatever other path BGP still has, or none.
Running it
Internet Exchange Point: 3 ISPs Peering opens on a canvas, already wired and addressed. Trace a packet through it hop by hop, or run the same topology live on real Linux network namespaces. How the two modes differ.
Open Internet Exchange Point: 3 ISPs Peering
Runs in your browser. No install, no signup, and the simulation starts instantly.
Open the lab →