Second, any application can run on any port. Rather than basing the decision solely upon the port number, the decision makers must consider the applications that will run on those ports.
For example, consider two different organizations running Web servers on port 80. One is an e-commerce site using the Web server to sell goods to the public. The other is a military intelligence organization using the Web server for internal sharing of highly classified information. In this black-and-white example, it's fairly obvious that the e-commerce site needs a firewall rule allowing port 80 from the Internet to the Web server. On the other hand, the military intelligence outfit would definitely not want to allow a similar kind of inbound access.
The best answer I can offer you is that you need to consider the risks and make a decision appropriate for your enterprise. Here are a few questions you should ask yourself that will help inform your decision:
- What is the business case for opening these ports?
- What service(s) will run on the exposed ports?
- What is the destination scope of the rule? Will it be limited to systems in the DMZ? Will it be limited to a single system or a broad group of systems?
- What is the source scope of the rule? Will it be limited to a single system or network, or will it be exposed to the entire Internet? Who controls those system(s)?
- What type of information will be transferred over this connection? Will it be encrypted?
For more information:
This was first published in January 2008