BGP

The current version, BGP-4, is defined in RFC 4271 and runs over TCP port 179. The internet consists of tens of thousands of autonomous systems, each with its own routing policy and identified by an AS number. BGP lets them announce to one another which IP prefixes are reachable via which paths. It is a path-vector protocol: every route carries the list of ASes it has traversed (AS_PATH), which also prevents loops.

Sessions between different ASes are called eBGP, sessions inside one AS iBGP. Best-path selection is steered with attributes such as LOCAL_PREF, AS_PATH length, MED and communities. This is how an ISP sends traffic to its preferred upstream, keeps a backup transit link, or peers directly with other networks at internet exchange points (IXPs).

An organisation that wants to connect to multiple upstreams (multihoming) needs its own AS number and IP block. Türkiye falls within the RIPE NCC service region for these resources. By the mid-2020s the full internet routing table had reached roughly one million IPv4 prefixes and several hundred thousand IPv6 prefixes, so routers that take a full table must have adequate memory and CPU.

A faulty BGP announcement can attract other networks' traffic (a route leak or hijack) and cause widespread outages. Mitigations include filtering inbound and outbound prefixes, max-prefix limits and route origin validation with RPKI (ROV, RFC 6811). MikroTik RouterOS v7 and carrier-grade routers support these mechanisms.