There are countless reasons for configuring routers with loopback interfaces. What is the best way to address the loopbacks? What sort of addressing scheme would produce the least amount of routing overhead, but still be scalable to grow to a 200-300 node network. I have heard of people allocating an entire class c address space per loopback, all the way down to addressing them with a 32 bit mask. What is the best way?
You can use whatever addressing policy you are following in your organization. Loopback is mainly used for some processes where an IP is needed for administrative purposes. Like your need to have a router-ID or an IP for IPSEC tunnel association.
A good practice is to have a block and assign IP addresses from that with /32 subnet mask. Care should be taken in assigning those IPs, because if you assign 192.168.1.1/32 to one router and 192.168.1.2 to another, and you have OSPF so 192.168.1.2 will become DR. So, much thought should be given before assigning any IP addresses to loopback interfaces.
Dig Deeper on Network Infrastructure
Related Q&A from Sudhanshu Gupta
One difference between managed and unmanaged switches is complexity. A managed switch is more complex and requires more skills, but it offers better ... Continue Reading
Unmanaged Linksys switches don't know where to send BootP and DHCP requests. Find out what to do to fix the problem of unmanaged switches and DHCP ... Continue Reading