Re: [Udp35] [Stackevo] Snowden and SPUD

[Michael Welzl <[e-mail]>]

> On 20. jul. 2015, at 20.42, Ken Calvert <[e-mail]> wrote:
> 
> 
>>> The axe-head attached to the stick is "if you don't let this through, then you get a giant pile of entropy that tells you nothing."
>> 
>> i heard that before and wondered, on wondered, why would anyone let any of that pass, then?
> 
> Same reason they would let QUIC pass: it will make applications work better - otherwise nobody will use it.  (Of course there's the obvious problem that it's not defined yet :-).  In any case, I agree that without noticeable improvement for users in some dimension(s), there's no point.
> 
>> so, and i'm serious: if we could do this with quic, that would help, right?
> 
> For some applications.  For others, it depends on whether the parts of QUIC "behind the veil" allow for evolution.
> Also, whatever is exposed will become ossified, so it should be designed with evolution in mind (whatever that means).

maybe that's not a problem?!

i just saw a part of brian's email that i had overlooked before as i was answering on my phone:

me:
>>> Now, assume that QUIC would become a general-purpose carrier. All kinds of applications would start using it. Everything would be multiplexed over QUIC. What would middleboxes do? They would start looking deeper into QUIC, right? Same problem again!

brian:
>> Except they can't, because QUIC encrypts everything except the things that are explicitly exposed to the middleboxes. This is the intention with SPUD as well -- we redraw the network/transport layer boundary, this time with a big fat red marker that says go ahead, bruteforce my keys.


... doesn't that indeed make SPUD the ideal carrier for anything?  People have an incentive to let it pass, already now - but given that it's encrypted, they can't really be sure it is web traffic. SPUD lacks the incentive, but QUIC has it - make QUIC your new SPUD and you have it???

Cheers,
Michael