hi Mirja, all,

Briefly, inline,

> On 28 Apr 2016, at 11:14, Mirja Kühlewind <[email protected]> 
> wrote:
> 
> Hi all,
> 
> another idea would be to provide some side channel to transport additional 
> information such as potentially options instead of encrypting the actual 
> options. In this case we would probably rather need a length field than a 
> single bit (and then we could use the same semantics as used for options to 
> fill this field...). Just an idea... probably need to further think about it.

Yep. I think the main win here, as David says below, would be protecting 
something a la HICCUPS, where hash information about the packet header would be 
tcpcrypt-protected, to allow endpoint detection of header manipulation.

> Mirja
> 
> 
> On 27.04.2016 20:24, David Mazieres wrote:
>> Mirja Kühlewind <[email protected]> writes:
>> 
>>> Hi all,
>>> 
>>> I briefly brought this up in the last meeting and would like to start
>>> the discussion on the mailing list now. The working group decided that
>>> tcpinc will not encrypt the TCP header for good reasons. However, it
>>> would still be possible to encrypt TCP options. This could help
>>> keeping confidentiality and would avoid that a middle could alter
>>> information in a option or strip it. Not sure if there is a case where
>>> some options should be encrypted and some not but I guess that would
>>> be possible as well. Any thoughts?
>> 
>> Encrypting existing TCP options strikes me as a bit complex, especially
>> since at that point you might to authenticate them.
>> 
>> That said, there's a related point which is that tcpcrypt, given its TLV
>> structure, could easily offer some sort of side channel.  In BA, Jana
>> suggested there might be good uses for something like this.  For
>> example, suppose you want to collect information on what middleboxes are
>> doing.  You could transmit an advisory copy of the options in a special
>> tcpcrypt frame and compare them at the receiving end.
>> 
>> We are definitely open to accommodating such scenarios for the simple
>> reason that it would be extremely simple.  In fact, we currently have 6
>> unused bits in the flags field of a frame.  We could easily say that the
>> high bit means this is a control message and implementations MUST skip
>> over control messages they don't understand.  (Or partition the space
>> using the other bits into some MANDATORY and a bunch of OPTIONAL
>> skippable ones.)
>> 
>> What do people think?

As an author of the SPUD requirements draft, this seems to me like a useful 
general purpose facility to have...

Cheers,

Brian


>> David
>> 
> 
> _______________________________________________
> Tcpinc mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/tcpinc

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc

Reply via email to