They blamed a configuration change gone wrong. If that was a change to their core routers, deployed to everything at once (or just to some crucial parts first,

They blamed a configuration change gone wrong. If that was a change to their core routers, deployed to everything at once (or just to some crucial parts first, even), it's quite believable that they could take (public-facing, at least) network down in one fell swoop.