The drivers for network programmability and software-defined networking are numerous and growing. Network infrastructure components represent one of the last elements of the IT service delivery chain that require manual provisioning.

Automating and orchestrating network infrastructures through programmatic interfaces can provide many benefits, including reduced provisioning time and increased service velocity. A more dynamic, agile, repeatable, and reliable approach to network device configuration, operation and monitoring is necessary to keep pace with the rapid change of business today.

Automation and orchestration methodologies that have been successfully applied to compute, virtualization, and storage platforms in the data center can also be applied to network fabrics. Perhaps the most impactful of these methodologies is the utilization of Application Programming Interfaces (APIs). When network platforms expose APIs as a means for configuration, control, and monitoring, it becomes possible to replace existing manual processes with automated workflows.

The extensibility of an API allows for network components to be controlled and managed in a centralized fashion. Some practical examples involving centralized management through APIs include:

• Network controllers leverage APIs to manage and operate the network inclusive of monitoring key performance indicators (KPI) such as latency, jitter and delay. The controllers' knowledge allows for routine change actions to be performed, such as dynamic traffic movement away from identified saturated links.
• Policy controllers leverage APIs to apply policies to meet the security and performance requirements of a given application, service, or user - all based on high-level policy definitions. This section will provide an overview of non-programmatic management interfaces available on Cisco Nexus switching platforms, as well as introducing the new open NX-OS programmatic interfaces for Cisco Nexus switches.