|
|
Posted Apr 05, 2009 at 7:38:41 PM
Subject: VPN for Server
Hi Guys,
I'm not new to Linux, but what I'm trying to do is going to be a bit more complicated than what I've ever attempted before.
I have an old Colbalt Raq P3 server coming my way, which for compatibility reasons will probably be running CentOS 4.x. My current Internet connection is NAT'd and behind a firewall, however I have a special server PPTP account that busts through the NAT firewall and gives me a static, public unlimited IP.
What I want to do, is configure this server to always connect to the PPTP connection, and bind all server software to the VPN-provided public IP.
I also want to use the Raq's 2nd ethernet port to run a NAT/Firewall for between 1 and 3 local machines. So, basically, the machine would be connecting to VPN, getting a real outside IP address and unrestricted tunnel connection, running server processes (apache) on the public, VPN provided IP, and also running a small Internet sharing NAT system from the VPN connection to the second Ethernet interface.
For those with doubts about the VPN working for this purpose, I have already tested this out on a Mac OS X desktop machine, and it seems to work fine for the small volume of data I'll be working with.
I don't have the server hardware yet, but I wanted to start prepping and reading documentation so I can make the setup as painless as possible when it arrives. Does anyone have any experience running a server over VPN, and providing a similar setup to what I'm trying to do?
Any advise or other help would be greatly appriciated!
|
Rubberman
Joined Jul 30, 2007 Posts: 944
Location:40 miles west of Chicago
Other Topics
|
Posted:
Apr 11, 2009 10:10:27 PM
Subject: VPN for Server
I would suggest CentOS 5.3 (latest release) if possible. There shouldn't be many compatibility issues unless you have some really proprietary software that is totally wedded to version 4. As for the server software running over a VPN, it should not be an issue. The servers shouldn't care where the connection is coming from. It will reply to the appropriate TCP/IP connection as that is a function of the fundamental network stack on the operating system. Assuming TCP (not UDP) connections, the servers will open and wait on a socket bound to a particular port. When it connects to a client it creates a connection socket and any further communication with that client takes place over that socket. The server doesn't know what network connection is being used. As I said, the low-level TCP/IP network stack takes care of all that cruft. Perhaps you should pick up Comer and Stevens' books "Internetworking with TCP/IP" (3 volume set). It's kind of the bible of TCP/IP design, implementation, and programming. I'm sure Amazon.com will be happy to help you part with your $$ and send them to you PDQ. :-)
Sometimes real fast is almost as good as real time.
Remember, Google is your friend!
|