Sun Nov 27, 2022 10:58 am
It is rarely a decision of "people out there" when configuring their WAN subnet - they have to accommodate to how the ISP has configured things on its end. The subnet size is a concern when the WAN address is a public one, because if you use the simplest method (direct assignment of the public address to the WAN on a point-to-multipoint type of interface), you waste 4 public addresses to provide a single public IP assigned to the customer - a /30 network contains a "network address", a "broadcast address", and two host addresses, one of which must represent the router at the ISP side. Even if you use a /31 subnet (which has a special RFC and is not directly supported by RouterOS), you still waste two addresses to deliver one. But this is rarely seen in reality, ISPs use larger subnets and port isolation, or use CGNAT addresses and route public ones via them rather than assigning the public ones directly to interfaces, or use PPPoE...
When your Mikrotik's WAN is connected to the LAN of the router provided by ISP to your home, this is only a concern if the ISP gives out public addresses on this LAN interface.
Large private networks are a different thing - there, the administrators may have to fit into a dedicated address range and partitioning it into smaller subnets may have some advantage for them.