An enterprise network operations team is replacing a legacy distributed routing environment with a controller-based architecture using Cisco DNA Center. When an administrator configures a new quality-of-service (QoS) policy across all campus switches, how does the deployment of this policy differ in the controller-based model compared to the traditional network model?
- The policy is defined once on the centralized controller and automatically provisioned to all managed switches using southbound APIs, whereas traditional networks require manual device-by-device configuration via CLI or SNMP.Answer
- BThe controller uses northbound APIs to directly modify hardware ASIC lookup tables on each switch, eliminating the need for local data plane forwarding hardware.
- CEach switch maintains its autonomous control plane and uses agentless SNMP traps over northbound interfaces to negotiate policy synchronization directly with peer switches.
- DThe local control plane on each switch is completely eliminated, forcing all network traffic data frames to be encapsulated into JSON structures and sent to the controller for hop-by-hop forwarding decisions.
Answer
In a controller-based network model, policies are defined centrally on the controller and provisioned down to network devices via southbound APIs, replacing per-device CLI manual configurations.
Controller-based networking abstracts network management and centralizes policy control. Administrators specify intent once on the controller (e.g., Cisco DNA Center), which uses southbound APIs to translate policies and configure all underlying network devices automatically. In contrast, traditional networks rely on per-device management where each switch must be configured individually via CLI or scripts.
Step-by-Step Solution
Key Concept
Centralized Management vs Distributed Configuration and Southbound API Roles
Estimated Time:1m 30s