Automation and Programmability

A team is choosing an automation tool that pushes configuration to network devices over SSH using human-readable YAML playbooks and runs without a permanent agent installed on each device. Which tool matches this agentless, push-based description?

• ATerraform, which uses HashiCorp Configuration Language to declare and provision infrastructure resources
• BA REST client alone, which issues raw HTTP verbs against device APIs but has no playbook engine
• CAnsible, which is agentless and pushes YAML playbooks to devices over SSH from a control nodecheck_circle Correct
• DNETCONF by itself, which is a configuration protocol carrying XML payloads rather than an automation tool

Why A is wrong: Terraform is a valid current tool but declares infrastructure in HCL and focuses on provisioning state, not pushing YAML playbooks over SSH, so the description does not fit.

Why B is wrong: A bare REST client sends HTTP requests but provides no YAML playbook automation framework, so it does not satisfy the stated workflow.

Why C is correct: Ansible is agentless and uses a push model, connecting from a control node over SSH and applying YAML playbooks, which matches every detail of the requirement.

Why D is wrong: NETCONF is a management protocol that exchanges XML over a secure transport, not an agentless tool that runs YAML playbooks, so it is the wrong category.

A platform engineer wants a declarative tool to provision cloud and network infrastructure, where they describe the end state and the tool tracks resources in a state file and works out the changes needed to reach it. Which tool fits this declarative, state-tracking model?

• ATerraform, which uses a declarative configuration and a state file to converge infrastructure to the described end statecheck_circle Correct
• BA purely imperative scripting approach where the engineer codes each step and ordering by hand with no resource tracking
• CAnsible used only in ad hoc mode, issuing one-off commands without any record of the resources it created
• DA SNMP polling manager that reads device counters and raises alerts but never provisions infrastructure

Why A is correct: Terraform takes a declarative description of desired infrastructure, records existing resources in a state file, and computes the plan needed to reach that state.

Why B is wrong: An imperative step-by-step script is the opposite of the declarative, state-tracked model the engineer wants, so it does not fit.

Why C is wrong: Ad hoc Ansible commands run procedurally and keep no resource state file, so although Ansible is a real automation tool, this usage does not provide the state tracking described.

Why D is wrong: An SNMP manager monitors devices rather than provisioning resources, so it cannot fulfil a declarative infrastructure-provisioning requirement.

An automation engineer is mapping the API tiers of a controller-based SDN solution. A custom orchestration application needs to request network services from the controller without knowing how the controller talks to the underlying switches. Which API does the application use?

• AA northbound API, typically a REST interface, that the application uses to consume controller servicescheck_circle Correct
• BA southbound API such as NETCONF or OpenFlow that the controller uses to program the devices
• CAn east-west API used to synchronise state between two peer controllers in a cluster
• DA southbound API such as RESTCONF that pushes configuration straight from the application to each switch

Why A is correct: Northbound APIs sit above the controller and let applications request abstracted network services, usually over REST, without dealing with device-level protocols.

Why B is wrong: Southbound APIs face the network devices below the controller, so an application asking the controller for services would not call them directly.

Why C is wrong: East-west APIs handle controller-to-controller communication for redundancy and scale, not an application requesting services from a single controller.

Why D is wrong: This bypasses the controller entirely and mislabels a device-facing protocol, whereas the application is meant to interact only with the controller through the northbound tier.

A REST client repeatedly polls router R1's API for interface counters. To avoid changing any state on R1 and to allow the response to be safely cached and retried, which property must the chosen HTTP method have, and which verb provides it?

• AIt must be safe and idempotent; GET reads the resource without altering server statecheck_circle Correct
• BIt must be idempotent but not safe; DELETE can be retried without altering the result
• CIt must create a fresh resource each call; POST guarantees a new snapshot every time
• DIt must fully replace the resource each call; PUT keeps the counters consistent

Why A is correct: GET is defined as both safe and idempotent, so repeated polls only read the counters and never change R1, which is exactly what safe caching and retry require.

Why B is wrong: DELETE is idempotent but not safe because it removes the resource, so using it to poll counters would destroy the interface object instead of reading it.

Why C is wrong: POST is neither safe nor idempotent and is intended to create resources, so repeated polls would attempt to create objects rather than harmlessly read counters.

Why D is wrong: PUT replaces the entire resource representation and thereby changes server state, so it is not safe for read-only polling even though it is idempotent.

While reviewing JSON returned by a device API, an engineer needs to identify which structure holds an ordered list of interface entries rather than a single set of named attributes. Which JSON construct represents that ordered list?

{
  "hostname": "core-sw1",
  "interfaces": [ "Gig0/1", "Gig0/2", "Gig0/3" ]
}

• AAn object, written with braces and made up of key and value pairs that are unordered
• BAn array, written with square brackets and holding an ordered sequence of valuescheck_circle Correct
• CA string, written inside double quotation marks to hold a single block of text
• DA boolean, written as the literal true or false to express a single condition

Why A is wrong: A JSON object uses braces and holds unordered key/value pairs, which describes a record of named attributes rather than an ordered list.

Why B is correct: A JSON array is delimited by square brackets and stores an ordered list of values, exactly matching the interfaces collection in the sample.

Why C is wrong: A string is one scalar text value in quotation marks, so it cannot represent a list of three separate interface entries.

Why D is wrong: A boolean carries only true or false and cannot hold an ordered collection of interface names.