[ih] Internet-history Digest, Vol 10, Issue 1

[Toerless Eckert]

On Mon, Jul 06, 2020 at 05:51:23PM -0400, John Levine wrote:
> >Probably too simplistic. Should not have been a reason for all those
> >E164 for companies tht are public knowledge (like their 'www' domain name).
> 
> A system that works only haslf the time is for most purposes worse
> than no system at all. Since they had to build some other system to
> deal with phone numbers that have privacy concerns, they might as well
> use that system for everything.

IPv6 addresses also come with privacy concerns, yet they are deployed
and workarounds are made for privacy. Besides, the privacy issues would
have been (IMHO) easy to resolve.

> >I would guess that the complexity of SIP/ICE/STUN/TURN for p2p connections
> >played a much more key role.
> 
> SIP for point to point works fine if you know the address.

Not if a side is behind NAT/FW. STUN/TURN/ICE exist for a reason.
Even with IPv6, multi-homing requires or benefits from some subset.

Lets see when we will get ubiquitous peer-to-peer media, end-to-end encrypted
without the rendezvous service seeing the media or knowing the keys.

In reality there is a lot more desire to perpass than privacy behind
what happened.

Cheers
    Toerless