This is a topic that is being discussed off the web among some computer programmers and I want to get some input for the next generation UPnP security implementations.
It was pretty much everyone has agreed on a mac address based control system so they system will only respond to the mac addresses in the list.
Do you think this is a better control over this than just ip assignment?
Granted, this is a feature us end users won’t see until 5 years from now, but I was asked to get feedback.
Another protocol just increases attack surface? I had a PC biz for 20 years and since I always disabled upnp on the router and never had an issue with things connecting I would question having it. But perhaps I am missing the need.
Mac address filtering is not adding security i think.
They are too easy to fake and often also messed up by wlan extenders. (They alter the mac address of the repeated packets) This is the reason that blue access sometimes fail.
I haven’t dissected blue’s network code thoroughly, but when I first started using IPFire, I did noticed it didn’t pass my penetration test but I haven’t revisited that since the fix was applied to encode the password. The blue failed to escape the inline code of the comma character, as well as the Perl CGI white space escape character when I tested 186. The fix was to use an access point and not use the auth mechanism.
Administrating UPnP via mac address has been a discussion for quite some time, as the demands of administrating control over the service has been in need since it was implemented.
As far as the repeaters changing mac address, further investigation into it seems to be needed. Because I wouldn’t think they would parse and change the original mac address in the packet as that would take some processing power to do that, but rather establish a bit offset or encapsulate in a AH style header like what ipsec does to the packet. Of course other mechanism exist they could be using like changing it to a L2 packet and the original L3 packet is inside of that.