Full network automation isn’t ready, but current tools can still help

As enterprises try to accelerate their digital transformation efforts, network agility is more important than ever, and network automation can help by ensuring quick access to data, maintaining network performance and improving security.

With work-at-home policies preventing IT teams from physically accessing most locations, automation is being used to perform remote updates, fix problems in data centers and branch offices, and to assist work-at-home efforts. It can also help eliminate human scripting errors to close configuration gaps that lead to security breaches.

While networking suppliers are advancing network automation tools and despite the talk of self-driving networks, IT professionals need to look to incremental gains in network productivity that current offerings can provide.

Benefits of network automation

Network automation tools can eliminate time-consuming manual tasks such as provisioning new networking gear, changing configurations or implementing new policies.  The benefits include:

• Reducing the hours needed to complete network-operations tasks
• Reducing human error that can cause outages
• Helping identify causes of network slowdowns
• Determining what devices are on the network

Broader benefits can include faster provisioning of new networks in an enterprise, improved network performance thanks to fewer configuration errors, and improved security by reducing configuration errors that create security vulnerabilities.

Evaluating automation

Given the immense variety of network architectures, there is no one-size-fits-all path for automating the network. IT teams need to ask themselves which processes, elements and locations are most likely to benefit from automation.  For example, what are the high-value, time-intensive tasks involved in current operations? Which processes can be automated quickly given available tools?

Automation tools can provide a single source of truth as IT teams inventory devices attached to their networks. Other tools can provide end-to-end visibility across endpoints, applications, network and cloud to help IT analyze performance of specific applications.

Most teams have alert fatigue, with large enterprises reporting thousands of alerts per day, so they could benefit from help understanding and making actionable the information that is available about the network. Automation tools should enable IT to better understand and fix network slowdowns, connection errors, managing quality of service, and capacity planning.

Network-operations options

Moving from fully manual to fully automated network operations is a long journey, and innovative IT teams choose to identify and tackle specific elements of automation. They might select one domain at a time – branch, campus or data center – on which to focus their efforts. Or they might pick a type of network equipment – Wi-Fi, routing, SD-WAN, switching – for which configuration management can be highly valuable.

There are three choices enterprises can make in deciding what automation tools to use:

• Rely on their primary network equipment supplier
• Consider point automation solutions from third-party network-management suppliers
• Do it themselves with open-source tools like Ansible.

Each option has pluses and minuses depending on the architecture of the network they are being considered for. And all three options can be combined to tackle individual automation challenges.

Networking and IT suppliers are promoting solutions that they claim leverage machine learning and artificial intelligence to “solve” network management challenges.  Machine learning is widely used today to gather useful information from large networking datasets and to suggest what issues to look for.

The complexity of most networks and related network security challenges mean that it is very unlikely that networks will “run themselves” any time soon. But currently available products may be adapted to situations where most networking tasks are still performed manually. New tools are available to enable incremental automation, and IT teams should evaluate aspects of their networks that can benefit from these new capabilities.