Simulation, Test and Automation Tools

What it is

Simulation and test tools let you build and check a network without risking a live one, and automation tools apply changes to real infrastructure.

Cisco Modeling Labs and pyATS are the simulation and test tools on the outline. Ansible, Terraform and Cisco NSO are the automation tools.

Used together, they let you design a change, test it in a lab, and then apply it to production the same way every time.

How it works

Each tool has a clear job, so match the tool to the stage of the work.

Cisco Modeling Labs

Cisco Modeling Labs, or CML, is a network simulation platform. You draw a topology of virtual routers, switches and other nodes, and run them as virtual machines that run real network operating system images.

The benefits are:

• A safe lab for learning and for testing a change before production.
• A topology that can be saved, shared and rebuilt on demand.
• A REST API and a Python client, so labs can be created and started by a script or by a CI pipeline.
• The ability to load a copy of a design to test a configuration against it.

CML suits proof of concept work and training. It does not replace a test in the real environment, because virtual nodes may behave differently from hardware.

pyATS and Genie

pyATS is a Python based test framework from Cisco. Genie is the library of parsers and models that sits on it.

It works from a testbed file, which is a YAML description of your devices, their addresses and how to log in.

devices:
  R1:
    os: iosxe
    type: router
    connections:
      cli:
        protocol: ssh
        ip: 192.0.2.1

Genie parsers turn the text output of a command into a Python dictionary, so a test can check structured data.

from genie.testbed import load

testbed = load("testbed.yaml")
device = testbed.devices["R1"]
device.connect()

output = device.parse("show ip interface brief")
print(output["interface"]["GigabitEthernet1"]["ip_address"])

The common use is to take a snapshot of the network before a change, make the change, take a second snapshot and compare the two. This shows exactly what the change affected. pyATS can also run as part of a CI pipeline.

Ansible

Ansible is an agentless automation tool. It connects over SSH or an API, so nothing has to be installed on the devices. You describe tasks in YAML playbooks, and the next pages show how to read them.

It is easy to learn and has many modules for servers and network devices.

Terraform

Terraform describes infrastructure in a declarative language and builds it through providers. You write what you want, run terraform plan to preview, and terraform apply to create or change it.

Terraform keeps a state file that records what it manages. It is strong for creating cloud and platform resources, and providers exist for Cisco platforms such as ACI and Meraki.

Cisco NSO

Network Services Orchestrator is a model driven orchestrator. It uses YANG models for devices and for services, and network element drivers to talk to many vendors.

Changes are transactional and can be rolled back. It suits service provisioning in large, multi vendor networks.

Comparing the automation tools

Where each tool fits

• Learn or test a design: CML.
• Check that a change did what it should: pyATS.
• Run configuration tasks across servers and devices: Ansible.
• Create and track cloud or platform resources: Terraform.
• Orchestrate services across many vendors: NSO.