An adjacent-network attack is a hostile action that only works when the attacker is already on the same local network segment, VLAN, wireless trust zone, or similarly close path to the target. It does not require direct exposure to the public internet. In practice, the attacker may need Layer 2 access, local routing reachability, or the ability to send traffic inside a trusted industrial subnet.
This matters because many systems expose discovery, management, or configuration services only to nearby hosts. Those services can still be abused for scanning, spoofing, denial of service, or protocol fuzzing if segmentation is weak or an insider/device on the segment is compromised. Defenses include strict network segmentation, limiting who can join the segment, disabling unnecessary local services, and monitoring for unusual traffic on protocols meant to stay local.