We've been developing a web application (PHP, using the Yii PHP framework) that is going to be used for data entry. The clients will be users from both the LAN and WAN (many of the remote clients will be behind a proxy, reaching our network using one IP address with NAT). What we basically want is to guarantee the validity of data in the way that no malicious user alters it.
Is there a way to programmatically identify each client in a unique way, so that I can guarantee (at least at some good percent) that no malicious remote user will connect? We were thinking of gathering the MAC addresses of all remote users and using a (non-web) client that generates a hash string that the user will input in the web application and then proceed if this authentication scheme passes. As I said, using other non-web applications for the remote client is an option.
Is such a solution as the one I describe above viable? Should we see other solutions, like maybe a VPN?