Security question:
If I call a phone number (say, the one on my debit card), I can be pretty sure that the person who answers actually owns that number.
On the other hand, if I _receive_ a phone call that purports to be from that phone number, I *can't* trust that the person on the other end actually controls the number (which is why it can often be a good idea to hang up and call back).
So here's the question: Why is the "to" information so much more reliable than the "from" info?
@codesections Because routers direct traffic towards its destination, but they don't care where it came from.
> Because routers direct traffic towards its destination, but they don't care where it came from.
Yeah, that makes 100% sense for email. When a router gets a message from A and needs to send it on to C, it only needs to have an accurate address for C—A could be anywhere.
But for a call (or other bidirectional communication) I don't understand how that works. Once the call starts, data has to go both ways. Doesn't the router need both to know A's location to send data from C?
@codesections Not necessarily; telephone networks don't follow the end-to-end principle, so the router can just pass return data back to the previous hop, which can do the same again etc.
In classical telephony, this was done implicitly by creating physical metal-to-metal connections, and there's still some of that left in local exchanges and probably other places.
@codesections One reason for maintaining this kind of connection-oriented routing could be performance; once a virtual circuit (as I believe the terminology is) is established, the router only has to check the circuit number against its virtual circuit table for each data packet, not repeat the whole routing process.
@alexbuzzbee @codesections Another reason for virtual circuit routing is that it makes bandwidth reservation simple.
