BGP is the protocol that stitches independent networks together. It’s how the internet — and any network with more than one provider or exit — decides which path a destination’s traffic takes.
What it is
Networks are grouped into autonomous systems (AS). BGP is how one autonomous system tells another which destinations it can reach, and by what route. Each advertised route carries attributes that determine which path wins:
- AS-path — the list of networks the route has crossed (also used to prevent loops).
- Local-preference — a way to prefer one exit over another; higher wins.
- MED — a hint to a neighbour about a preferred entry point; lower wins.
- Communities — tags used to drive routing policy.
- Next-hop — where traffic is actually sent.
A router advertises the destinations it can reach to its neighbours and receives destinations from them. The chosen path is the outcome of comparing all of those attributes.
Why it matters
BGP is powerful, but it is also fragile and — for most monitoring tools — invisible. When traffic quietly shifts to a slower or more expensive path, the cause is almost always a BGP attribute change: a dropped local-preference, a lengthened AS-path after a session reset, or a re-tagged community. Interfaces stay “up,” dashboards stay green, and yet performance degrades or connectivity breaks. A single misconfiguration or an upstream provider’s mistake can reroute or black-hole your traffic in seconds.
What good looks like
Real BGP assurance means seeing not just that a route exists, but why it was chosen — the full set of attributes on the routes you advertise and receive — and keeping a history so you can compare today against yesterday. When a path changes, you want to know immediately which prefix moved, what attribute changed, and when. That turns a multi-hour investigation into a few minutes.