A regional utility runs OSPF between substations over private MPLS and a multipoint Ethernet service linking three control centers that keeps electing an unstable DR. Security audit requires authentication of all OSPF exchanges. Which OSPF authentication and network type design should the designer select for stability and security?

Cryptographic authentication deployed on all OSPF interfaces satisfies the audit requirement by providing keyed integrity and replay protection for hellos and all link-state exchanges across MPLS and Ethernet, while point-to-multipoint on the three-site Ethernet eliminates DR/BDR election by treating the multipoint fabric as a collection of point-to-point adjacencies, which directly resolves instability from DR flapping and simplifies flooding without static peering. The design retaining broadcast with tuned priorities fails by stability concept because priorities only influence who wins election but preserve the DR state machine, so any flap still triggers re-election and adjacency churn. The design using NBMA with static neighbors fails by stability and operations concept because NBMA still elects a DR and depends on manual unicast neighbor statements, which is fragile and ill-suited to multicast-capable Ethernet. The design imposing point-to-point with full mesh fails by topology concept because point-to-point expects exactly two routers per link, so it cannot natively model a three-center multipoint service and forces extra subnets and adjacencies. Exam caveat is to distinguish authentication scope from area scope and to recall that mismatched authentication or incompatible network types prevent FULL adjacency, with message-digest versus HMAC-SHA variants tested explicitly. Practical check is to verify ip ospf authentication message-digest with matching key and ip ospf network point-to-multipoint on the Ethernet interfaces, then confirm with show ip ospf interface showing no DR and show ip ospf neighbor showing stable FULL states.