Mailing List Archive

Rekeying
Hello,
I apologize that this is slightly off topic.

According to the Internet Draft I found for SSH ver 1
(draft-ietf-tls-ssh-00.txt from Jun 13, 1996), the client or server can
send a SSH_MSG_KEXINIT at any time to force a new key exchange. I looked
through the code for OpenSSH and ssh-1.2.27 and can't find where it does
this. I then searched the Secure Shell mailing list archives and saw some
comments that rekeying will be implemented in Ver 2.

Did I miss the code in OpenSSH? Are there any ver. 1 implementations that
provide rekeying? The draft also says that the session key is broken up
to provide different encryption keys and iv's, but the code seems to use
the same keys for both directions and iv = 0. Is there anywhere that I
can get a more up to date protocol spec?

thanks,
brian
Re: Rekeying [ In reply to ]
Hello Brian,

1) draft-ietf-tls-ssh-00.txt is obsolete. There is no known implementation.
2) OpenSSH-1.2.2 and ssh-1.2.27 implement the SSH-1.5 protocol.
This version does not support rekeying or IV's other then zero.
The file RFC.nroff from the OpenSSH distribution describes SSH-1.3.
3) SSH2 supports all the features you mentioned (IVs, different
keys and algorithms per direction, rekeying and much more).
See the draft-ietf-secsh-* series for more information, especially
draft-ietf-secsh-transport-06.txt

-markus

On Tue, Feb 15, 2000 at 04:22:19PM -0500, Brian Carrier wrote:
> According to the Internet Draft I found for SSH ver 1
> (draft-ietf-tls-ssh-00.txt from Jun 13, 1996), the client or server can
> send a SSH_MSG_KEXINIT at any time to force a new key exchange. I looked
> through the code for OpenSSH and ssh-1.2.27 and can't find where it does
> this. I then searched the Secure Shell mailing list archives and saw some
> comments that rekeying will be implemented in Ver 2.
>
> Did I miss the code in OpenSSH? Are there any ver. 1 implementations that
> provide rekeying? The draft also says that the session key is broken up
> to provide different encryption keys and iv's, but the code seems to use
> the same keys for both directions and iv = 0. Is there anywhere that I
> can get a more up to date protocol spec?
>
> thanks,
> brian
>
>
>
>