From: <ch...@to...> - 2004-03-23 19:30:11
|
sorry Clemmitt I Just re: to you first time meant to re: all Disclaimer: IANAMCSE and IANAcLD If the bridging is setup correctly then only the ethernet level of the stack should be used for the unfiltered data because windows does not have an address on the physical lan all TCP/IP trafic will be filtered. But Clemmitt's point is valid because there could be non TCP/IP holes such as netbeui or any other network layer protocols that are left open on windows or a carefully formed TCP/IP packet that could exploit a flaw in the lower layers and also any raw ethernet holes that there may be in windows. Long story short this would not be any better than any other host based firewall except for maby a lot more flexability. Other windows specific host based firewalls probably take the fact that they are running on windows and at least holler at you if netbeui or something else is on. chris > Hi, > > On Tue, 23 Mar 2004, Hugo Campos wrote: >> Hi, i've been playing with CoLinux and i've been thinking about how one >> would configure a windows-colinux with a single public IP so that the >> public IP was assigned to colinux. Then it (colinux) would share (NAT) >> internet access to windows. That would effectivelly provide a Linux >> Firewall, which can be configured with fun things like packet shaping >> and all, in a Windows box with (i believe) little overhead for normal >> DSL/Cable speeds. > > I think Chris's earlier post sounds like something neat to try. But > I'll repeat what I said on the Help forum FWIW. > > Disclaimer: IANAcLD. Windows is the host OS for coLinux. To access the > coLinux network functionality, the packets have to pass through the > Windows network stack first. So I don't think it's possible for a coLinux > installation to filter the incoming packets for the instance of Windows > it's running on. > > But I think Chris has a good point. If what you want is packet shaping > or filtering of Windows user traffic, his idea sounds like an excellent > place to start. I don't think this will protect against an attack > that exploits network-related holes in Windows. > > If I've got this wrong, someone please correct me! > > Clemmitt > > |