Internal security of Rev?

Dar Scott dsc at
Wed Jul 12 14:24:49 EDT 2006

On Jul 12, 2006, at 2:27 AM, John Tregea wrote:

> Yes my original question was about protecting classified  
> information within a database where the front end may end up being  
> Rev based.

You are mixing two levels.

If the information is encrypted and decrypted with a "hardwired" key,  
it is virtually part of the application.  It is subject to the  
limitations discussed.

However, if it is encrypted using a user supplied key, then it stands  
alone as encrypted information.  The user supplied key might be a  
certificate, a passphrase, or a passphrase to get to a certificate.

You can mix the user key with "hardwired" key to slow down decrypting  
without the user supplied key.

So, you need a user-supplied key.

> But the structure of the stand alone rev application is my  
> remaining concern. (unless you all think of some stuff I haven't  
> thought of.)
> The classified information would specifically be for supply chain  
> risk assessments under ISO 28000 and 28001. We hope to use Rev to  
> build a front end to a proprietary database structure, but have to  
> know we can certify the resulting application under ISO 17799  
> (Information Security Management) before clients would be prepared  
> to use the product/service.

Revolution uses good cryptographic functions.  It uses a library that  
has undergone review and has a controlled build distribution process.

However, 1) you don't know what kinds of sneaking things folks at  
RunRev have put into their code.  I don't think they have done  
anything sneaky, I mean you are not able to demonstrate that they  
have not, without going to extra effort.  Also, 2) I would not be  
surprised if there are RAM info leaks through Revolution's memory  
management.  That is, unused memory might be returned to the system  
without being written over.  Rev does a lot of copying.  (I do some  
things to mitigate this, but I have no idea if they really do any  
good.)  If those are not a concern for ISO 17799, I think you are OK.

Rev encryption is based on openSSL libraries and you need to make  
sure you have a good copy of a reviewed version.  Don't ship with  
what you get from RunRev for Windows; download a new copy and check  
the digest.

SSL is normally app to app, so you should be OK there, too.  However,  
at this time, you cannot supply a cert from the client side with  
Rev.  (Unless something happened when I was asleep.)  Hmmm.  There  
should be an enhancement request for this.

Dar Scott

More information about the Use-livecode mailing list