Your question is unclear but I'll try to answer the best I can, within my understanding.
- Do you want to control machines from connecting to any port on your machine? if so, you need to control the built-in windows firewall or find yourself a filter driver you can control. In order to write your own filter driver, you must leave the land of managed code, so I am guessing that's not an option.
To learn how to control the firewall, here's a link:
http://www.shafqatahmed.com/2008/01/controlling-win.html
more on google.
- Do you want to control remote machines from connection to a port on your machines that your application owns? You cannot do that either (see #1 above). However you can take action after the connection, and close the connection if you don't like the remote IP (check the remote endpoint's IP).
two caveats with this approach:
It doesn't save you from a DoS attack.
You will need to be careful if you need ipv6 support (you can't just check the IPV4 address in that case)
HTH