Re: multiplexing -- don't do it

On 04/03/2012 07:32 AM, Mike Belshe wrote:
> On Tue, Apr 3, 2012 at 1:30 AM, Stephen Farrell
> <stephen.farrell@cs.tcd.ie>wrote:
>
>>
>>
>> On 04/03/2012 12:22 AM, Robert Collins wrote:
>>
>>> This seems rather timely:
>>>
>>> http://www.telegraph.co.uk/**technology/news/9179087/**
>>> Internet-activity-to-be-**monitored-under-new-laws.html<http://www.telegraph.co.uk/technology/news/9179087/Internet-activity-to-be-monitored-under-new-laws.html>
>>>
>>
>> And not timely but relevant:
>>
>> http://tools.ietf.org/html/**rfc2804<http://tools.ietf.org/html/rfc2804>
>
>
> So from my reading of this, what we're discussing as a https trusted proxy
> does not classify under RFC 2804's definition of wiretapping in any way.

Probably. That'll depend on what goes in an I-D I guess.

But yes, while 2804 is relevant for the Daily Telegraph link above,
I would hope its not relevant for HTTP/2.0.

S.

>
> Section 3 defines 4 definitions of wiretapping, and we are not hitting any
> of them:
>
> 1) Not wiretapping because what we're talking about here is explicitly
> known to the sender.
>
> 2) Not wiretapping because the receivers would see that this was done by a
> proxy user agent, in the same way that a standard proxy today identifies
> itself in the middle (note, however, that transparent proxies today are
> considered wiretaps by this definition!  so by removing transparent
> proxies, we're removing a common class of wiretap!)
>
> 3) Not wiretapping because the sender can be told of the exact consequence
> and also be able to configure out of it.
>
> 4) Not applicable - we aren't proposing this at all.
>
> Any disagreement?
>
> Mike
>
>
>
>>
>>
>> S
>>
>>   -Rob
>>>
>>>
>>>
>>
>

Received on Tuesday, 3 April 2012 09:30:59 UTC