I am trying to (more or less) uniquely identify a system for licensing purposes. I have chosen the computer’s on-board network adapter’s MAC address for this task, since I can be sure that every cmputer running this software actually has one, and this avoids re-activation when changing e.g. the harddrive.
I am having troubles reliably identifying the onboard network adapter, though.
Using the “Win32_NetworkAdapterConfiguration” ManagementClass, I can get a whole lot of MAC Addresses, including the address I like, but I have not found a way to distinguish the onboard one from virtual adapters installed by Windows or Virus Scanners.
This list seems to be ordered, though. The MAC Address I am interested in is (on my machine) listed before other (real) network adapters. (The list is ordered by interface index.)
Using NetworkInterface.GetAllNetworkInterfaces(), I think I can identify the real network adapters by filtering on .NetworkInterfaceType == NetworkInterfaceType.Ethernet, but this list seems to be unordered (an added network card appears before the onboards one).
Is first using the second method to get a list of real networks cards and then sorting them by the order of appearence in the first list a reliable way of identifying the MAC address I am looking for? Can the interface index in the first list change?
I’d be happy to hear your thoughts!
Thanks!
P.S.: I know that the MAC address can be rather easily changed, but I can live with that. I cannot live with the customer not being able to use the software after simply inserting a WLAN stick =)
A rather low-tech solution would be to invoke the netstat command and look for the MAC address of the adapter that has a valid IP address. I’ve never seen the netstat command fail on a machine whereas I’ve seen WMI give unexpected results numerous times.
In any case, I have done a similar activation system before and I used the MAC address as the identifying key. In the end, it wound up being more trouble than it was worth – both for me and the customer! What I found to be a far better balance and less hassle was to have the user “sign in” the first time the software was installed. With the user’s consent, you could send some piece of identification to the server such as their MAC address.
Then you only need to periodically check your activation database for evidence of major license violations and deactivate the keys as necessary. As a customer that hates product activation, and an ISV that hates software piracy, I can see both sides of the argument and this way it avoids putting the customer in the uncomfortable position of having to convince you they are legit when something (inevitably) goes wrong.
Just to name a few reasons why MAC identification may not work… I use two NIC’s (wired and wireless) in my laptop depending on whether I’m at work or home. One or the other may be disabled at any time. The other thing to note is that I use virtual machines quite a bit and not only do they get their own MAC but I could specify any MAC I want. Then of course one day you’ll find out that you have like 100 people in your database with a MAC of all zeros. 🙂 Nothing is guaranteed here.