Re: [TLS] assert TLSext in renego-ServerHello instead of disable renego

[Yoav Nir <[e-mail]>]

Actually, there have been messages suggesting that clients sending this (or any other) extension, may have their session terminated, especially if using SSLv3 and even with some TLS 1.0 servers. If that is true for even a small minority of servers (say, 1%) we can't realistically expect browser vendors to add this, except buried deeply in some advanced menu.

I wish we had some hard data on this, but if this is true, we may need to find other ways for the client to signal support, or to signal re-negotiation.
________________________________________
From: [e-mail] [[e-mail]] On Behalf Of Michael D'Errico [[e-mail]]
Sent: Friday, November 13, 2009 18:10
To: [e-mail]
Subject: Re: [TLS] assert TLSext in renego-ServerHello instead of disable renego

>> My point was that if the server *is* upgraded and the client does not
>> send the extension on its initial handshake, it is taking an unnecessary
>> risk.
>
> The issue isn't whether the client sends it. Of course the client
> sends it. It's whether the client aborts the connection of the server
> doesn't return it.

I'm in agreement with you that the client should send the extension.  But
there have been messages suggesting that a client doesn't need to send it
on its initial connection, only when renegotiating.  I've been trying to
convince them that the client must send it especially in its first handshake
in order to avoid being attacked.

Mike
_______________________________________________
TLS mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/tls