There are two philosophies to blocking ports, and which one is appropriate for your network depends on the level of security that you require.
For a high-security network, especially when storing or maintaining confidential data, it is normally recommended to "filter by permission." This is the scheme in which all ports and IP address permissions are blocked, except for what is explicitly required for network functions. For instance, port 80 for web traffic and 110/25 for SMTP can be allowed to come from a dedicated address, while all other ports and addresses can be disabled.
Most networks will enjoy an acceptable level of security by using a "filter by rejection" scheme. When using this filtering policy, ports that are not used by your network and are commonly used for Trojan Horses or reconnaissance can be blocked to increase the security of your network. For instance, blocking ports 139 and 445 (TCP and UDP) will make your network more difficult to enumerate, and blocking port 31337 (TCP and UDP) will make you more secure from Back Orifice.
This should be determined during the network planning phase, when the level of security required is compared to the needs of the network users.
Fortifying router security
Step 1: Change the default password!
Step 2: Disable IP directed broadcasts
Step 3: Disable HTTP configuration for the router, if possible
Step 4: Block ICMP ping requests
Step 5: Disable IP source routing
Step 6: Determine your packet filtering needs
Step 7: Establish Ingress and Egress address filtering policies
Step 8: Maintain physical security of the router
Step 9: Take the time to review the security logs
About the author
Chris Cox is a network administrator for the United States Army, based in Fort Irwin, California.
This tip originally appeared on SearchNetworking.com.