Re: [TLS] COMMENT: draft-ietf-tls-renegotiation

[Huang Min <[e-mail]>]

 
Hi, Stefan

I don't agree the point that unpatched renegotiation could be still used and
all the problems solved by applications.
The vunerability is brought by SSL/TLS protocol itself, it MUST solved by
itslef. We could not give this protocol to other upper applications,and say:
"hi, here is a protocol you could use, but there is a vunerability in the
protocol which can be used for a MITM attack, so you SHOULD do A, B, C,
D.....to avoid it, if you want to use this protocol. " It is not
responsible. 

Discarding renegotiation mechanism or fixing the vunerability are both more
acceptable than leaving this problem to upper application layer. 

regards,
Min Huang


-----邮件原件-----


Date: Tue, 15 Dec 2009 02:37:01 +0100
From: Stefan Santesson <[e-mail]>
Subject: Re: [TLS] COMMENT: draft-ietf-tls-renegotiation
To: Nicolas Williams <[e-mail]>,	Russ Housley
	<[e-mail]>
Cc: [e-mail], [e-mail]
Message-ID: <[e-mail]>
Content-Type: text/plain;	charset="US-ASCII"

The fact that it is used for legitimate purposes makes me worried about the
consequences of just abandoning it without fixing it.

Perhaps a middle way is to not fix it and not to remove it, but simply
clarify it's security properties. I'm still not convinced you can't use
unpatched renegotiation securely if applications simply don't accept
messages sent before completed renegotiation as being delivered under the
context of the renegotiated connection. Other's have suggested the same.

/Stefan