試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 58 / 65
Which tactic is most appropriate for minimizing user response latency while keeping costs down?
①Maintain model instances of the same scale in all regions at all times.
②Centralize all inference in a single central region.
③Combine edge caching and lightweight local inference.
④Abandon real-time processing and transition completely to batch processing.
⑤Always perform synchronous replication between regions.