Re: RMTCMD's security?

[Scott Klement <midrange-l@xxxxxxxxxxxxxxxx>]

Hi Gary,

I certainly did not say that securing what can be executed on a system 
is silly!

Securing SBMRMTCMD won't stop a worm from going here or there -- THAT is 
my point.

If there were only two systems in the world, and SBMRMTCMD was the only 
possible conceivable command to submit a command, then your philosophy 
might work...   but unfortunately, that isn't reality.

An analogy:

Securing the server against running remote commands (which is what I 
advocate) is like putting your money in a safe so a robber can't break 
into your house and steal it.

Securing the server by disabling the client command (which is what you 
seem to be advocating) is like going to every conceivable criminal's 
house and breaking their legs so they can't travel to your house to 
steal your money.


On 5/17/2012 4:59 PM, Monnier, Gary wrote:
> Hi Scott,
>
> I'm surprised you feel securing what can be executed on a system as silly.  Not doing so, from my knothole onto the world, seems absurd.  To me, it is akin to buying a padlock for a locker but not actually locking it!!
>
> If you can't get from system A to System B because the stuff allowing it is unavailable to you who cares how tightly system B is locked down?  System A isn't letting you get there anyway.
>
> A worm goes here
> A worm goes there
> Until it is stopped from going anywhere.
> It is stopped when it cannot execute stuff allowing it go somewhere.