On 3/15/2015 3:04 PM, Olivier Bonaventure wrote:
> Joe,
> 
>> ...
>>>> Protecting against spoofed RST would be a very good thing indeed.
>>>> But we want a solution that is robust in the presence of
>>>> middle-boxes, and there is a generic issue there.
>> ...
>>>> Joe, do you have a design that solves that?
>>
>> Actually, to be more direct:
>>
>> Don't use a NAT.
>>
>> when you're behind a NAT, you're beholden to how the NAT interacts with
>> the Internet, and *you* are on a private communications system that is
>> decidedly NOT the Internet.
> 
> I agree that the Internet would be easier without NATs, but looking at
> the TCPINC charter, the first requirement for TCPINC is :
> 
> "The high-level requirements for the protocol for providing TCP
> unauthenticated encryption and integrity protection are:
> - It should work over the vast majority of paths that unmodified TCP
> works over, in particular it must be compatible with NATs (at the very
> minimum with the NATs that comply with BEHAVE requirements as
> documented in RFC4787, RFC5382 and RFC5508)."
> 
> 
> This requirement has been met by Multipath TCP. TCPINC should also aim
> at meeting this requirement.

Sure, but there will never be a NAT that prevents an off-path RST attack
from DOSing a connection. Once you accept that NATs have control over
your connection but you don't have control over NATs, you're done with
claiming protection from NAT-directed attacks.

That's true for ALL TCP approaches though NATs.

Joe

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

Reply via email to