milanuk
08-21-2001, 07:25 PM
This is a subject I see brought up from time to time on Usenet, mailing lists, etc. But either I loose the thread before it gets a good answer, or I missed it all together. The scenario is something like this:
A LAN, of whatever size, that uses a DHCP server to assign client IP addresses, feed info on DNS servers, whatever. Perhaps at a university dorm complex, or an apartment complex, or a business hotel that wants to offer high-speed internet jacks in the rooms. So they _have_ to let people randomly connect to the network, query the DHCP server, and subsequently, connect to the Internet.
The problem isn't w/ the honest people, the ones who pay for their access, and use it accordingly. The problem is w/ the ones who decide that thru a little shoulder surfing, they can access the net using their neighbors machine name. Or perhaps that they already know, either from a friend or whatever, the necessary info like name server addresses, gateway address, etc. They go right ahead and connect to the Internet, w/ relative impunity.
I've seen this topic come up for OpenBSD, Linux, hell, even WinNT 4.0. How do you limit the access out the firewall/gateway/router to legitimate accounts w/ valid IP's from the DHCP server? I've seen people mention doing something w/ the MAC addresses, assigning 'permanent' DHCP leases based on the MAC address, therefore requiring users to have previously registered their NIC's MAC address w/ the business office. But how do you tie that in to the firewall? It seems like it should be extremely easy, on first glance, but on second & third looks it seems more and more of a major PITA.
Any ideas or suggestions out there?
Monte
A LAN, of whatever size, that uses a DHCP server to assign client IP addresses, feed info on DNS servers, whatever. Perhaps at a university dorm complex, or an apartment complex, or a business hotel that wants to offer high-speed internet jacks in the rooms. So they _have_ to let people randomly connect to the network, query the DHCP server, and subsequently, connect to the Internet.
The problem isn't w/ the honest people, the ones who pay for their access, and use it accordingly. The problem is w/ the ones who decide that thru a little shoulder surfing, they can access the net using their neighbors machine name. Or perhaps that they already know, either from a friend or whatever, the necessary info like name server addresses, gateway address, etc. They go right ahead and connect to the Internet, w/ relative impunity.
I've seen this topic come up for OpenBSD, Linux, hell, even WinNT 4.0. How do you limit the access out the firewall/gateway/router to legitimate accounts w/ valid IP's from the DHCP server? I've seen people mention doing something w/ the MAC addresses, assigning 'permanent' DHCP leases based on the MAC address, therefore requiring users to have previously registered their NIC's MAC address w/ the business office. But how do you tie that in to the firewall? It seems like it should be extremely easy, on first glance, but on second & third looks it seems more and more of a major PITA.
Any ideas or suggestions out there?
Monte