G
Guest
I'll try to keep this simple
I've got a multi-homed Win2K server with one NIC connected to the internet, and two other NICs connected to different local subnets. The goal is to allow the local subnets to access the internet, but not each other. I also need to support VPN. Since I have to use RRAS for VPN, I am using NAT with IP forwarding enabled (instead of ICS) to allow the local subnets to access the internet
So everything is working great, except the local subnets can now talk to each other because IP forwarding is enabled to support NAT. For technical reasons, I can't use packet filtering in RRAS to block this traffic. Which leaves IPSec
I set up an IPSec filter that blocks all packets going from subnet A to subnet B (using the "A specific IP Subnet" option for both source and destination). When I apply this filter, traffic between subnets is _not_ blocked. If I change the filter so that the destination is "Any IP Address", traffic from the source subnet is blocked. Unfortunately, it is blocked _everywhere_, so now I can't get to the internet. If I instead change the filter so the destination is "A specific IP Address", and enter the address of a host on subnet B, all traffic from subnet A destined for subnet B _is_ properly blocked. But going this route will require me to enter a filter for every host on both subnets (yuck!)
So, my question is: is this behavior by design? At first I thought maybe IP forwarding was short-circuiting something and bypassing the IPSec filters, but it appears that it's only "broken" for subnet-to-subnet filters. Any ideas
Thanks in advance
- Chris
I've got a multi-homed Win2K server with one NIC connected to the internet, and two other NICs connected to different local subnets. The goal is to allow the local subnets to access the internet, but not each other. I also need to support VPN. Since I have to use RRAS for VPN, I am using NAT with IP forwarding enabled (instead of ICS) to allow the local subnets to access the internet
So everything is working great, except the local subnets can now talk to each other because IP forwarding is enabled to support NAT. For technical reasons, I can't use packet filtering in RRAS to block this traffic. Which leaves IPSec
I set up an IPSec filter that blocks all packets going from subnet A to subnet B (using the "A specific IP Subnet" option for both source and destination). When I apply this filter, traffic between subnets is _not_ blocked. If I change the filter so that the destination is "Any IP Address", traffic from the source subnet is blocked. Unfortunately, it is blocked _everywhere_, so now I can't get to the internet. If I instead change the filter so the destination is "A specific IP Address", and enter the address of a host on subnet B, all traffic from subnet A destined for subnet B _is_ properly blocked. But going this route will require me to enter a filter for every host on both subnets (yuck!)
So, my question is: is this behavior by design? At first I thought maybe IP forwarding was short-circuiting something and bypassing the IPSec filters, but it appears that it's only "broken" for subnet-to-subnet filters. Any ideas
Thanks in advance
- Chris