![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
I've now learned quite a bit about the BSD firewall system. I'm impressed.
Take a machine -- we will call it A -- and throw in three Ethernet NICs. A1, A2 and A3. Hook A1 up to the ISPs wire (RED zone) and A2 up a hub that is in the DMZ or green zone.
Do not assign IP numbers to either of these cards. Create an electronic bridge between the two NICs. Set up rules for packet filtering and routing in both directions.
These rules can be very complex or very simple, depending on needs. But one can allow only certain kinds of traffic to reach certain machines in the Green zone.
In the meantime, nobody can access machine A because it has no addresses.
Now, take another machine -- call it B -- and put three NICs in it, too. B1, B2 and B3. All these NICs have addresses. B1 is connected to the Green hub or DMZ and has an address reachable from the Internet. B will do NATing and further firewall actions for the most secure Blue zone. So B2 is on an internal only TCP net. 192.168.x.x or 10.10.x.x. B1 and B2 are bridged the same way as A1 and A2.
Two firewall machines. Each filtering and routing packets according to the grand scheme of the network designer.
Machine A can also route packets to an off site machine or to an on site machine that's sole purpose is to entice crackers and script kiddies. They sniff around this honey pot machine until the security gooroo (who has been notified by email) can start tracing. Cloak and Dagger time. The honey pot can look like anything, but most often it looks like a Microsoft server with sensitive information like credit card numbers and IDs.
And what of A3 and B3 NICs? Well, A3 is set up to receive it's IP by dhcp from B through B3. That way, the only way to control A is through the more secure B. If one is outside the system, one has to get through two firewalls and know two different passwords before one can affect the operation of A.
This is all too cool.
Some time ago, I read an article that really blew my socks off. When you shut down a linux machine, services are shut off and file systems are unmounted. One can no longer access or change the state of the kernel or run any applications. But the kernel is still running. suppose one did not shut off the network services? One could have a kernel with IP filtering and routing going on that is totally impossible to access in any manner at all.
The only problem I see with it, though, is how would one monitor it to see if it is still performing it's job properly?
Doesn't matter though. The BSD system I've been playing with is great fun and almost as secure.
Take a machine -- we will call it A -- and throw in three Ethernet NICs. A1, A2 and A3. Hook A1 up to the ISPs wire (RED zone) and A2 up a hub that is in the DMZ or green zone.
Do not assign IP numbers to either of these cards. Create an electronic bridge between the two NICs. Set up rules for packet filtering and routing in both directions.
These rules can be very complex or very simple, depending on needs. But one can allow only certain kinds of traffic to reach certain machines in the Green zone.
In the meantime, nobody can access machine A because it has no addresses.
Now, take another machine -- call it B -- and put three NICs in it, too. B1, B2 and B3. All these NICs have addresses. B1 is connected to the Green hub or DMZ and has an address reachable from the Internet. B will do NATing and further firewall actions for the most secure Blue zone. So B2 is on an internal only TCP net. 192.168.x.x or 10.10.x.x. B1 and B2 are bridged the same way as A1 and A2.
Two firewall machines. Each filtering and routing packets according to the grand scheme of the network designer.
Machine A can also route packets to an off site machine or to an on site machine that's sole purpose is to entice crackers and script kiddies. They sniff around this honey pot machine until the security gooroo (who has been notified by email) can start tracing. Cloak and Dagger time. The honey pot can look like anything, but most often it looks like a Microsoft server with sensitive information like credit card numbers and IDs.
And what of A3 and B3 NICs? Well, A3 is set up to receive it's IP by dhcp from B through B3. That way, the only way to control A is through the more secure B. If one is outside the system, one has to get through two firewalls and know two different passwords before one can affect the operation of A.
This is all too cool.
Some time ago, I read an article that really blew my socks off. When you shut down a linux machine, services are shut off and file systems are unmounted. One can no longer access or change the state of the kernel or run any applications. But the kernel is still running. suppose one did not shut off the network services? One could have a kernel with IP filtering and routing going on that is totally impossible to access in any manner at all.
The only problem I see with it, though, is how would one monitor it to see if it is still performing it's job properly?
Doesn't matter though. The BSD system I've been playing with is great fun and almost as secure.
no subject
Yes, I confess I'm finding understanding this stuff pretty neat too!