Re: iSeries FTP security

["Brad Stone" <brad@xxxxxxxxxxxx>]

On Wed, 18 May 2005 15:51:41 -0500
 Patrick Botz <botz@xxxxxxxxxx> wrote:
> > the ftp exit does raise the security level to a certain
> degree
> 
> I disagree with this statement -- with or without
> cannonicalization
> problems.

I don't.. 

> 
> Exit point programs allow you more flexibility in
> how/when/who might access
> something, but that is not the same as saying that they
> provide additional
> security.  The security comes from your access control,
> and the flexibility
> comes from the exit point.

This is a reach.  

How (It's ftp)
When (exit points with access times, good idea, I may have
to add that to my FTP app)
Who (user profiles)  

I don't see how exit programs provide flexibility...
 unless you've got hooks in your programs to run processes
after certain commands are executed.

> 
> Access control can protect you against any failings of an
> exit point
> program.  But exit point programs cannot protect you
> against failings of
> your access control model. 

If everyone has *ALLOBJ, you can use FTP exit points to
control what users have access to using FTP and what FTP
commands they can use.  

When referring to FTP alone this does provide security.  

This doesn't mean you shouldn't have  good object security
in place, though.  It may be wrong to have a poor object
authority plan, but I'd say more do than don't.  :)

And even if you your exit
> point program works
> for FTP, what about those interfaces that don't have exit
> points?

Now that's thinking outside the box to prove your point.

I repeat this again.  Why isn't IBM's FTP exit point
consistenet when it comes to reporting paths?  It resolves
most paths to qualified paths.. but not cannonical paths.

Brad