FYI: SSH1 now disabled at compile-time by default

[Dan Kaminsky]

Alright, so I pulled the data from scans.io,  There's actually 82,650
devices on the open Internet claiming support for <=SSH-1.5, generally
routers.  Top 20 on that is:

$ head -n 20 ssh1_versions.txt
  39148 SSH-1.5-Cisco-1.25
  14477 SSH-1.5-HUAWEI-VRP3.1
  10571 SSH-1.5-1.0.0
   4634 SSH-1.5-HUAWEI-VRP-3.10
   3284 SSH-1.5-1.2.33
   2965 SSH-1.5-VRP-3.3
   1836 SSH-1.5-VRP-3.4
   1125 SSH-1.5-Server
   1034 SSH-1.5-X
   1022 SSH-1.5-1.2.22j4rad
    753 SSH-1.5-1.2.26
    538 SSH-1.5-OpenSSH_4.4
    422 SSH-1.5-SSH
    379 SSH-1.5-HUAWEI-VRP5.0
    304 SSH-1.5-OpenSSH_2.3.0_Mikrotik_v2.9
    270 SSH-1.5-SSHServer
    269 SSH-1.5-1.2.27
    223 SSH-1.5-OpenSSH_3.7.1p2
    175 SSH-1.5-Huawei
    136 SSH-1.5-LtSSH_3.6

Of course, this is out of a total of 15,124,618 SSH servers, granting a
compat rate greater than 99.99%.  However, distinctly and painfully unlike
SSL/TLS, SSH is successfully deployed and used on internal networks that
cannot be scanned from the open Internet.  It's also a protocol of fairly
critical importance, uniquely used in a "hop by hop" manner in which each
hop actually has to work.

7.3% of Cisco routers on the open Internet only support SSHv1.  The numbers
inside private networks are likely to be higher.

I can see the argument for pushing people to upgrade, but not by surprise
in a minor version.  If SSH is going to block old insecure versions it has
a much bigger problem, because upgrade rates on SSH on the Internet are
actually not fantastic.  Here's the top 40 across all versions of SSH:

$ head -n 40 sshall_versions.txt
2412684 SSH-2.0-OpenSSH_5.3
 984056 SSH-2.0-OpenSSH_4.3
 936855 SSH-2.0-dropbear_0.51
 854624 SSH-2.0-dropbear_0.46
 798414 SSH-2.0-OpenSSH_6.0p1
 790303 SSH-2.0-OpenSSH_6.6.1p1
 771396 SSH-2.0-OpenSSH_5.9p1
 465647 SSH-2.0-OpenSSH_5.5p1
 430372 SSH-2.0-ROSSSH
 338577 SSH-1.99-Cisco-1.25
 337282 SSH-2.0-OpenSSH_6.2
 325681 SSH-2.0-dropbear_0.52
 315864 SSH-2.0-dropbear_2012.55
 305623 SSH-2.0-dropbear_2013.58
 178282 SSH-2.0-OpenSSH
 169613 SSH-2.0-OpenSSH_4.7
 162892 SSH-2.0-OpenSSH_5.1p1
 160225 SSH-2.0-OpenSSH_5.3p1
 150068 SSH-2.0-Cisco-1.25
 141795 SSH-2.0-OpenSSH_5.1
 130653 SSH-2.0-OpenSSH_6.7
 122612 SSH-2.0-RomSShell_4.31
 121542 SSH-2.0-homepl
 116630 SSH-2.0-OpenSSH_5.8
 112202 SSH-2.0-OpenSSH_6.4
  98829 SSH-2.0-dropbear_2011.54
  96457 SSH-2.0-OpenSSH_4.3-HipServ
  88471 SSH-2.0-OpenSSH_5.2
  88444 SSH-2.0-OpenSSH_5.9
  86001 SSH-2.0-dropbear_0.48
  69497 SSH-2.0-OpenSSH_5.4p1
  67565 SSH-1.99-DOPRA-1.5
  60749 SSH-2.0-OpenSSH_6.1
  59323 SSH-2.0-ARRIS_0.50
  59103 SSH-2.0-lancom
  58497 SSH-2.0-Trendchip_0.1
  57757 SSH-2.0-OpenSSH_5.6
  55951 SSH-2.0-OpenSSH_6.2p2
  54750 SSH-2.0-dropbear_0.50
  52685 SSH-2.0-OpenSSH_6.6p2-hpn14v4

This is specifically a scenario where OpenSSH should measure twice and cut
once.  The worst case scenario is that people update even less than they
already do, because who knows what servers are no longer important enough
to require connectivity to next.  Start the discussion, absolutely, but
let's not forget even SSHv1 is miles better than Telnet, and this just
isn't the rapidly updating consumer browser ecosystem we're messing with
here.

--Dan




On Tue, Mar 24, 2015 at 9:16 PM, Damien Miller <djm at mindrot.org> wrote:

> On Tue, 24 Mar 2015, Nico Kadel-Garcia wrote:
>
> > If it's disabled by default in any major distributions, it's going
> > to break a lot of git repositories, svn+ssh repositories, and rsync
> > environments.
>
> The chronology doesn't support this.
>
> When Subversion 1.0 was released in 2004, OpenSSH had been defaulting to
> protocol v.2 for almost three years.
>
> git was first released two years later, at which time v.2 had been the
> default for over five years.
>
> Seriously, protocol 2 became the default in *2001* and the old protocol
> has been disabled for new sshd installs for the last eight years. If
> anything, we've moved way too slowly.
>
> > Can it wait until version 7, instead of being slipped into a minor
> > update?
>
> Our version number has been a simple counter for years; the first digit
> of the version has no significance beyond that.
>
> Distributions will make their own decisions about what to support and
> they already ship far more intrusive changes than flipping a configure
> switch. I hope they ship a "openssh-ssh1" package instead of just
> setting it back through.
>
> -d
>