On Thu, Jan 06, 2005 at 04:07:15PM -0800, Casey Schaufler wrote:
--- Klaus Weidner <klaus(a)atsec.com> wrote:
> I had made the implicit assumption of independent administrative
> domains as was done in the previous Linux security targets, but there
> are of course other ways to define this. I would still consider
> passwordless rsh to be an abomination though ;-)
Even on a cluster?
Hmmm, this is getting a bit off-topic. If it's a sufficiently secure
network (such as a cluster where the network is not exposed to
potentially hostile users) I guess it would be okay, especially if it's
mainly used for automated administrative tasks and not directly by users.
The reason I'm opposed to it is mostly due to the temptation for misuse
over networks that are not secure when people are used to this type of
infrastructure. For typical interactive sessions the performance impact
of encryption is not a big issue, and I think it's preferable to train
people to always use encrypted protocols rather than needing to explain
the scenarios where unencrypted ones are safe or not.
-Klaus