Krebs is not wrong but he also doesn't get into how many things are "verified" through an even older mechanism, the LOA (letter of authorization/letter of autho

Krebs is not wrong but he also doesn't get into how many things are "verified" through an even older mechanism, the LOA (letter of authorization/letter of authority), which is literally just an actual letter on one corporate entity's letterhead, specifying some CIDR-notation prefixes/ranges that are permitted to be announced by that entity's upstreams, signed by a corporate officer. Often in the form of a raster scanned PDF. BGP4 in general dates back to a time when everybody trusted each other and the global routing table/network engineering community was a much smaller thing. We've been trying to glue better things on top of it for 25+ years.