General: What is a dynamic gateway, how does it work, where is it used?

Question

I saw this option in my OPNsense gateway configuration, and was hoping someone could shed light on what it is, exactly:

Dynamic gateway policy:  [YES/NO]
This interface does not require an intermediate system to act as a gateway

With this explanation:

If the destination is directly reachable via an interface requiring no intermediary system to act as a gateway, you can select this option which allows dynamic gateways to be created without direct target addresses. Some tunnel types support this.

I found this googling and it is getting me a little closer to understanding: https://kb.fortinet.com/kb/documentLink.do?externalID=FD48900#

Once enabled, dynamic-gateway hides the gateway variable for a dynamic interface, such as a DHCP or PPPoE interface. When the interface connects or disconnects, the corresponding routing entries are updated to reflect the change.

But I still don't quite understand. Would someone be nice enough to explain to me a real-world situation in which this might be used, and the difference between using a dynamic gateway or an intermediary gateway?

Answer 1

If the destination is directly reachable via an interface requiring no intermediary system to act as a gateway, you can select this option which allows dynamic gateways to be created without direct target addresses. Some tunnel types support this.

There is no single definition for "dynamic gateway", so the actual meaning depends on context. Here, OPNsense apparently talks about a link that doesn't require addressing to the next hop. That is typical for a point-to-point link (P2P) - a serial line, a tunnel, or similar - that doesn't require addressing because everything transmitted is received by the node on the other side. There is no "intermediate system" that would require addressing the gateway (like the data link layer of an Ethernet network).

Apparently, OPNsense uses the common "next hop address" paradigm as a general standard and needs to cope with the problem when there is no such address, so it creates a "dynamic" gateway internally as a workaround. On other systems, P2P interfaces can act as gateways without a next hop address.

Answer 2

I hope picture above shows it really nice, basically once one Gateway node for example GW1 will fail then server nodes will automatically use other dynamic gateways GW2 and GW3. Also dynamic gateways allows servers to load balance the traffic among them so you can avoid overloading one Gateway.