Re: [multipathtcp] comments on draft-paasch-mptcp-lowoverhead and draft-paasch-mptcp-ssl

[Christoph Paasch <[e-mail]>]

Hi Yoshifumi,

thanks for your comments. Please find my replies below.


On Thursday 25 October 2012 01:51:58 Yoshifumi Nishida wrote:
> 1: I'm wondering if experimental status might be better for them. Is there
> any thoughts on this?

You are probably right. I can change this for the next version.

> 2: How is the relationships between these drafts? Is it totally
> independent?

Yes, they are independent from each other. MPTCP v0 or MPTCP v1 (low overhead) 
can use a key provided by the application.

> 3: In my feeling, it could be dangerous If token is used for high-order
> 32bit. (draft-paasch-mptcp-lowoverhead)
>     We might want to emphasize this point.

What part in the draft are you referring to?
It is the random number that is used for the 32 high-order bits of the IDSN.

Or do you mean that we just should explain why the random number is used, and 
not the token for the high-order bits.

> 4: In section 5 of draft-paasch-mptcp-lowoverhead.
>     "if an attacker manages to join an existing connection...".
>     Does this mean the attacker steals the token? I just would like to
> confirm..

The attacker might know (through sniffing) the token or guess it and thus can 
add a subflow to the connection.


Cheers,
Christoph


-- 
IP Networking Lab --- http://inl.info.ucl.ac.be
MultiPath TCP in the Linux Kernel --- http://mptcp.info.ucl.ac.be
Université Catholique de Louvain
--