You need to enable a passthrough of whatever port you decide to use for LimeWire. It's not critical, and can be almost any port in the range between 1024 and 60000. Ports below that are reserved for system functions, and ports above 60000 are not normally used.

Since the RIAA has been reported to monitor the traffic on 6346, and some ISPs have been known to limit or block traffic on the default port, I'd use something other than 6346. You could use 6347, if it's easy to remember.

However, I still see the majority of people using the default port, so unless it's blocked, you can certainly use that one.

Basically, you want to forward this port on your firewall to the machine you are going to be using for your client. You will want it to still look like the port is at the firewall address, but you forward any requests to the client machine. So to the network, your client machine looks like the firewall.

Only TCP needs to be forwarded. You can still block UPD requests if you want, because the protocol does not use UDP at all.

I use a Linksys BEFSR firewall (now well under $60 at most computer stores). It has a forwarding feature that makes this very easy. You just enter the port number, and the IP of the local client machine, select TCP as the protocol, and you are done.

But a software firewall should work in the same basic way, as long as it has support for port forwarding. If it doesn't, then you want to find a way to let all requests for your specific port be passed through. That will work fine, but you need to insure that any other machines on the network do not have listeners enabled for that port, or they could be accessed too.

Since you can choose the port number from a large range, it should be possible to do what you want with almost any firewall software.

Hope this helps,
Rockkeys