Screen scraping versus API automation

For a long period of time, the CLI was the only access method available to manage and operate network devices. Operators and administrators used to have SSH and Telnet to access the network terminal for configuration and troubleshooting. Python, or any programming language, has two approaches to communicating with devices. The first one is to use SSH or telnet the same as before and get the information, then process it. This method is called screen scraping and requires libraries that will be able to establish a connection to the device and execute a command directly on the terminal, and other libraries to process the returned information to extract useful data from it. This method often requires knowledge of additional parsing languages, such as regular expressions, to match the data pattern from the output and extract useful data from it.

The second method is called an Application Programmable Interface (API) and this method depends entirely on sending a structured request using REST or SOAP protocols to the device and returning the output, also in structured format, encoded in JSON or XML. The time needed for processing the returned data in this method is quite small compared to the first method; however, the API requires additional configuration on network devices to support it.