What Is the Difference Between Network Automation and Orchestration?

(This Tech Primer is sponsored by Itential.)

With the complexity of network environments on the rise, there is more pressure than ever to deliver operational excellence in the networking domain to keep up with the rapid pace of modern applications.

More data, more apps, and more hybrid environments means that networking staff can no longer depend on manual activities to keep up with the needs of their infrastructure. They are turning to automation and orchestration approaches such as infrastructure as code (IaC), platform engineering, and NetDevOps to deliver smooth-running operations.

In this Tech Primer, we’ll take a look at the various network automation and orchestration technologies and how they differ.

Network Automation vs. Orchestration

The world of technology buzzwords can be confusing, with people speaking different languages. To set things straight, we’ll define the differences between network automation and network orchestration and explain how they are changing.

On a basic level, think of network automation as the narrower of the two fields, implementing software to execute repetitive tasks or scripts. For example, in a software-defined wide-area networking (SD-WAN) environment, an SD-WAN platform might be able to execute network automation by configuring hundreds or even thousands of devices using templates, and then it validates the configurations.

Other network automation tasks might include running Python scripts or Ansible playbooks to automate repetitive tasks. For example, some network automation tasks might include the following:

• Configuration changes
• Device provisioning
• Configuration checks
• Performance checks
• Security scans

One challenge is that many of these network automation tasks might be limited to specific vendor platforms or technology domains. For example, the automation might be limited to an SD-WAN vendor's products. If you want a wider range of automation that extends across systems or domains—or possibly incorporates business logic to respond to specific events—that's when we enter the world of orchestration, which is more sophisticated in its programmability. Orchestration also enables automation across software or business environments (IT, cloud, or enterprise applications).

According to a recent Market Guide from Gartner, orchestration is more complex and difficult to implement. “Orchestration remains underutilized by I&O teams due to perceived complexity, assumed loss of control and steep skills requirements,” says the Market Guide.

Tying Automation Together with Orchestration

As networking teams implement solutions for network orchestration, this can take their network automation to the next level.

Network orchestration requires more sophisticated tools to tie automation together across domains. This has spawned the IaC or NetDevOps movements, which use integration approaches and technology tools such application programing interfaces (APIs) to drive automation across technology domains and silos. IaC and NetDevOps can build bridges to network orchestration by tying together a variety of elements such as network monitoring, CI/CD pipelines, and application code to build orchestration.

Orchestration needs to cross boundaries, for example from cloud to IT systems or business logic, to be able to program and configure infrastructure. As the Gartner Market Guide describes, the perceived complexity of full orchestration has held some teams back in its implementation.

Simplifying Orchestration

When complexity is climbing, the way to manage it is by creating abstractions and centralized resources that can tame the complexity.

Part of the challenge of integrating systems across diverse domains and systems is that different tools are used. A team might want to build standardized systems for driving orchestration across the entire infrastructure. This approach is behind the movement known as platform engineering, which aims to create consistent tools for teams to use to manage infrastructure, while at the same time remaining flexible (new tools can be added at any time).

An ideal approach to orchestration can enhance network automation by tying together network sources of truth, monitoring systems, and AIOps to both provide and output the right information to the right place. It can also enable workflows across network domains by using and improving the use of existing automation assets like Python scripts or Ansible Playbooks.

For more information, check out this guide to Network Automation vs. Orchestration from Itential.