From: Leslie Viljoen <leslieviljoen@...>
Date: 2006-05-23T22:23:05+09:00
Subject: Re: Writing a interpreter extension

On 5/23/06, Guest <krisleech@interkonect.com> wrote:
> Tim Becker wrote:
> >> How do those banks prevent the keys being found by their own staff or
> >> server admins?
> >
> > That's a different problem, typically solved by using secure hardware.
> > The point is, only the key needs to be secret, not the implementation.
> >
> >    -tim
>
> Agreed, but how do you hide the key from people with access to the
> server?

You should read up on how SSL works:
http://www.ourshop.com/resources/ssl.html

The keys are not embedded in the program but are secured using the
operating system. Private keys (certificates) are "signed" (encrypted)
by a third party known to be reliable by both bank and client. Each
private key is just a file on a harddrive, protected by the operating
system. Someone who had physical access to that harddrive can get the
key (if they can bypass the operating system) but we assume the bank
keeps it physically secure.

This is pretty similar to my SSH example at the beginning - keys are
held in files apart from the source code, secured by the OS. With SSH
you can also have a "key to a key" where the key file is itself
encrypted and the user has to type a password to decrypt it before it
can be used.

Les