Fellow monks! Lets not get confused between using public key authentication to enter a machine and having no passwords. The public key authentication is just as secure as the originating account. Keep in mind that in order to ssh into the remote machine the account on the far end needs to have the originating user's identity.pub in their authorized_keys file. Rather than using passphrases, you're using key exchange as authentication. Its still considered very secure. If any other user attempts to ssh into the remote machine, they will still be prompted for a username/passphrase pair. The risk you run is letting someone else get your public key. However, protecting it is fairly easy and no less complicated than choosing a good password that you don't keep under your mousepad on a sticky note. You may want to check out Univ. of Michigan's documentation on exactly what we're talking about and the risks associated with it.

As far as Rexec goes, I dont believe that Rexec is encrypted traffic, which, if we're nitpicking, would cause most infosec people to gasp out loud. Sure it can be encrypted just like rsync using ssh as the transport. So why not just use ssh to begin with?

To the original poster, if you need the user logging in to the remote box to issue superuser level commands, you might want to look into using sudo. Here is a pretty good and complete guide on using it. Once you get into this arena, you will also need to be careful about what sudo permission you allow the account that ssh's into the machine, so you're back to square one when it comes to planning your security measures, but so goes the business.

I read that Security and Convenience are inversely proportional. You'll have to be the one to decide where you draw the line at convenience and sleeping well at night.

humbly -c