I assume you already have a separate firewall.
Gordon Lee wrote:
> We have a single physical star ethernet network with a Cisco
> router connecting to the Internet. We have divided our
> physical network into two logical networks
> (mask=255.255.255.128).
It appears you want to use a subnet half the size of standard. I have not tried this so
I am not sure it will work with e-smith. I would use standard class C subnets just to
avoid uncertainty. There is no shortage of Private addresses!
> Should we use just one ethernet card with IP-aliasing, or use
> both ethernet cards, to achieve our purpose of internal
> routing?
I would use two ethernet cards.
The ip-aliases trick would save an ethernet card at the price of performance so I would not recommend it.
> And what is the procedure? Please give me as much
> details as possible because I am a novice with TCP/IP, Linux
> and e-smith. Thanks in advance!
Set up e-smith as 'server only' in console.
One ethernet card would connect to your cisco router and its subnet, the other to your second subnet. When you set up each ethernet card, make choices for your subnets and netmask that do not conflict with the standard e-smith setup. For example, if you use DHCP, make appropriate changes to the range.
The e-smith manager will allow you to connect to a second subnet with the second ethernet card. However, you may have to do some command line configuration to use smaller subnets. You would definitely have to for ip-aliases.
You mentioned 'logical networks' so I would suggest that the two networks be physically separate; otherwise you may still have a performance problem because of the network traffic. e-smith will route between the networks for your normal traffic but your backup data won't be seen on the 'other' network.
I hope I understood your problem. Let me know if this is helpful.
Regards,
Paul