When multiple routes exist to the same destination IP address, which rule do routers use...

The primary and most critical rule used by Junos OS (and routers in general) to determine the next hop for a specific packet is the Longest Prefix Match (LPM) algorithm. When the Routing Engine or Packet Forwarding Engine looks up a destination IP address in the routing or forwarding table, it may find several entries that technically encompass that address. The LPM rule dictates that the router must select the most specific route available—that is, the entry with the highest number of matching bits in its subnet mask (the longest prefix).

For example, if a router has a route for 10.1.1.0/24 and another for 10.1.1.0/28, a packet destined for 10.1.1.1 will always be forwarded according to the /28 route, as it is more specific. Longest prefix match takes precedence over all other selection criteria, including route preference (administrative distance) and metrics. Preference and metrics are only evaluated when the router has multiple entries for the exact same prefix length (e.g., two different paths to 10.1.1.0/24). This logic ensures that traffic is guided along the most precise path defined in the network topology. Routers never choose paths at random or based on the most recent update as their primary selection mechanism, as doing so would result in non-deterministic and inefficient routing behavior. Reference: Routing Fundamentals, Routing Table and Forwarding Table Selection.

==========