Source routing for a Windows server with 2 network interfaces and 2 public IP's on different networks...
Hi, I have a Windows 2008R2 Web Edition server connected to internet via two different ISP, meaning two network interfaces, two public IP's and... two different gateway on two different networks... the purpose is to reach the server from the outside using any of the two IP's, if one isp fail the server must remain available using the other one... Thus, i need to define a routing different according the source IP used... any traffic coming from the ip of the first ISP must use the gateway of the first ISP, and any traffic coming from the ip of the second ISP must use the gateway of the second ISP... My problem : i've found some informations on this subject, and it seems that the only solution involve the use of the Routing and Remote access... but it's not available on a Web Server Edition... My question : is there any other solution with Windows Web Edition (KISS if possible) ? For information, here's how i do with a Linux Box (if someone could point me to an equivalent solution with Windows, 100.100.100. = ISP1, 200.200.200. = ISP2) : ip rule add from 100.100.100.2 table ISP1 ip route add default via 100.100.100.1 dev eth1 table ISP1 ip rule add from 200.200.200.2 table ISP2 ip route add default via 200.200.200.1 dev eth2 table ISP2 Thanks in advance for any tips...
June 22nd, 2010 9:50pm

RRAS won't do it either.Bill
Free Windows Admin Tool Kit Click here and download it now
June 23rd, 2010 2:40am

Hi, Thank you for your post here. From the description, you want to achieve the ISP internet networking load balance and ISP internet connection fault tolerance to publish a web server. Please understand that RRAS (Windows Server) will not do what you want either. When a external user attempt to access the web site via FQDN such as www.yourdomain.com, how can it be resolved to the different public IP address you have? DNS Round bobbin out there may helps on this. However, please understand that Round bobbin will resolve the FQDN to IP address in a round mechanically. It will not detect whether one of the ISP connection is down or not.I think you will need a dedicated device/service to achieve the web server publishing.
June 23rd, 2010 8:46am

No, it's not what i want, forget the load balancing and fault tolerance, that's a totaly another subject, my question is much more "basic"... i just want to be able to use two IP's on two different networks with two different gateway, one for each IP... and i want to be sure that no trafic from IP1 goes trough gateway2 and no trafic from IP2 goes trough gateway1... in a perfect world i would also like to be able to do that with only one NIC, normaly with a basic source routing config like in my example it's not a problem, but this criteria is not realy critical. For the usage, some websites will use IP1, some others will use IP2, and i want to be able to reach the server for administrative task (RDP admin) using one or another IP. Currently, if i configure the two IP and the two gateway (one per nic), first i receive a big warning from windows telling me that they will not function proprely, then i have a routing table with two default routes with no source from restriction, meaning i cannot be sure that no internal routing will occur (trafic from IP1 routed trough nic2 and gateway2). When i look my routing table (route print), i see two routes with the same metric and no rules avoiding the internal routing... Thats my problem because of course, ISP1 does not allow me to send traffic coming from the IP of another ISP (ISP2), and thus i must enforce my configuration to avoid that, any packet with a wrong IP could result in a spoofing protection : the isp shut the port down...
Free Windows Admin Tool Kit Click here and download it now
June 23rd, 2010 9:28pm

No answer, does it mean that the WebServer Edition does not allow to use two different IP's from two different ISP ? :(
June 27th, 2010 10:20am

my friend, i think that you can do the same here, on the first interface you can put do the basic configuration i.e: IP Address, Subnet Mask, Defualt Gatway, ..... on the second interface use the basic configuration without Gateway this is becuase you can not put 2 gateways on the same server, to work around this you can use route add command and change Metric, i.e make a defualt route which is route add 0.0.0.0 mask 0.0.0.0 (Gateway) METRIC 2 by this way you can do it. Samer F. Mustafa Sr. Microsoft Platform System Engineer sf_mustafa@hotmail.com
Free Windows Admin Tool Kit Click here and download it now
June 27th, 2010 10:26am

my friend, i think that you can do the same here, on the first interface you can put do the basic configuration i.e: IP Address, Subnet Mask, Defualt Gatway, ..... on the second interface use the basic configuration without Gateway this is becuase you can not put 2 gateways on the same server, to work around this you can use route add command and change Metric, i.e make a defualt route which is route add 0.0.0.0 mask 0.0.0.0 (Gateway) METRIC 2 by this way you can do it. Samer F. Mustafa Sr. Microsoft Platform System Engineer sf_mustafa@hotmail.com but, wouldn't doing this cause it to try to push Both blocks out the interfaces. allow me to clarify: ISP 1: Server IP 1.1.1.2 GW .1 ISP 2: Server IP 1.1.2.2 GW .1 doing as you suggested, wouldnt it cause it to try to push both blocks out on ISP 1? and then if ISP 1 went down, push it out ISP 2? or am i misunderstanding/missing something?
January 27th, 2011 12:04pm

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics