Skip to content
  • There are no suggestions because the search field is empty.

The Evolution of Network Configuration

Programmability and automation are vital across IT deployments, driving a shift away from the CLI towards automation-friendly interactions with network, server, and storage devices. While configuration automation isn't new, modern large-scale networks demand effective management systems. 

As networks evolve to support numerous devices, protocols, and technologies, and as more dynamic provisioning methods become necessary, traditional CLI and SNMP management approaches gradually fade due to their limitations.

The next generation of configuration management requires interfaces with client-side validation, separation of configuration and operational data, built-in backup and restore capabilities, human and machine-friendliness, support for various transports, data encoding formats, efficient tooling, extensibility, openness, and a model-driven approach.

Model-Driven Programmability 

Next-generation network management relies on a programmatic and standards-based approach, leveraging data models to define configuration and state information in a uniform language. This allows for configuration automation across different Cisco devices and enables scalability, even though CLI configurations may be more user-friendly. 

Model-driven programmability streamlines configuration and control while supporting end-to-end service delivery, especially in Cloud Computing scenarios. The key elements of the comprehensive device API encompass the following components:

  • Data Models: Data models in networking provide standardized schemas, ensuring consistent representation of data across platforms. They don't send information directly to devices but rely on protocols like NETCONF and RESTCONF to communicate JSON- and XML-encoded documents adhering to a given model. These models define data object attributes, such as valid VLAN IDs, VLAN name conventions, and data types, ensuring configuration accuracy and standardization. The industry is transitioning from CLI-based configurations to fully modeled devices represented in JSON and XML, driven by robust models like YANG. Data models also encompass operational data and actions for network configuration transactions.
  • Transport: Model-driven APIs support various transport methods, such as SSH, TLS, and HTTP(s), enabling flexible communication.
  • Encoding: Various encoding formats enable applications to communicate with APIs efficiently. These formats, like XML, JSON, GPBs, and YAML, structure data for machine understanding while remaining human-readable. Understanding supported data types is crucial when using APIs to configure devices like Cisco routers. You compose requests in these formats, which an API server translates into router instructions, enabling specific configurations and actions. 
  • Protocols: Model-driven APIs offer protocol options, including NETCONF, RESTCONF, and gRPC. These protocols facilitate communication with network devices based on data models, while REST may also be used in specific cases. Protocol selection depends on networking expertise, programming preferences, and available tools.