BGP Fundamentals

How the internet stitches independently run networks together, and why that job needs a protocol built for policy and scale rather than shortest path.

What an autonomous system is, how AS numbers are allocated, and the difference between public and private 16- and 32-bit ASNs.

Why BGP is a path-vector protocol - carrying the whole AS_PATH for loop prevention, policy, and internet-scale routing.

How two BGP routers form a session over TCP/179, the four message types they exchange, and the finite state machine that takes them to Established.

Install Docker and Containerlab so you can run the hands-on labs on your own machine.

Explore a pre-built, already-Established eBGP session: read its state, FSM, timers, capabilities, and the BGP table - no configuration required.

The same BGP behaves differently depending on whether its two peers sit in different autonomous systems or the same one, and that single fact drives AS_PATH handling, next-hop, loop prevention, and where you peer.

BGP advertises nothing on its own; learn the two ways to originate your prefixes into the BGP table and when to reach for each.

Bring up an eBGP session between two routers in different autonomous systems.

Originate each router's loopback into BGP with a network statement and watch the prefix cross the eBGP session.

Build eBGP across three ASes from a blank slate so every AS reaches every other through a transit AS in the middle.

An AS has many routers, but only the edges learn external routes over eBGP. iBGP carries those prefixes across the whole AS, with a split-horizon rule standing in for AS_PATH loop prevention.

Why every iBGP speaker must peer with every other one, the n(n-1)/2 session explosion that follows, and a first glimpse of how route reflectors and confederations escape it.

Why iBGP sessions ride on loopback addresses instead of physical interfaces, how to source them correctly, and why an IGP underlay is the thing that makes loopback peering possible.

Bring up an iBGP session between two routers in one AS, peering on loopbacks over an OSPF underlay.

Diagnose an external prefix that arrives over iBGP but stays INVALID, then fix it with next-hop-self.

Build a complete transit AS from a blank slate: an OSPF underlay, a full iBGP mesh on loopbacks, next-hop-self on the edges, and two eBGP edges, so an external prefix crosses the AS end to end.

How BGP attaches metadata to every prefix in an UPDATE, and the four attribute categories that decide whether a value crosses an AS boundary or stays home.

How AS_PATH drives loop prevention and path length, and why a reachable NEXT_HOP is a precondition for any route to become best.

Use LOCAL_PREF to steer your AS's outbound traffic and MED to hint a neighbor AS about your preferred inbound entry point.

Weight is a Cisco/FRR local-only knob that wins first on one router; ORIGIN is a well-known mandatory attribute that breaks ties near the bottom.

Walk BGP's ordered tie-breakers from WEIGHT down to lowest neighbor IP to see exactly how a router picks one best path per prefix.

Raise LOCAL_PREF on routes from a preferred upstream to steer your AS's outbound traffic onto it.

Make one of two upstreams the preferred inbound path by prepending your own ASN on the advertisement you send the other.

Use MED across two parallel links to tell a neighbor AS which entry path to prefer, then walk the lower tie-breakers.

From a blank-slate edge router, engineer both outbound (LOCAL_PREF) and inbound (AS_PATH prepend) traffic so a dual-homed AS prefers one upstream over the other.

Write the policy that decides which prefixes BGP accepts and announces, using ip prefix-list with le/ge range matching.

Learn how route-map clauses match, set, and permit or deny BGP routes, and why the trailing permit clause is the difference between policy and a silent blackhole.

Use BGP communities to color routes with policy intent and act on that color elsewhere in your AS or at your upstream.

Accept only a customer's real allocations by applying an inbound prefix-list on an eBGP session.

Tag advertised prefixes with BGP communities and watch a downstream AS act on the tags, including no-export blocking propagation.

From a blank slate, build the inbound policy for a small ISP: filter a customer's bogon leak, tag accepted routes with a community, and raise their LOCAL_PREF so they win.