Go back

The Next Generation of Network Automation

February 21, 2022

Clearly, network automation is essential for organizations to keep up with the accelerating pace of technology and business. However, there are numerous tasks that current automation tools cannot facilitate. Over the last two decades, two types of network automation have evolved into mainstream practices to replace some of the manual work for network engineers.

  • In-house scripting is an ad-hoc approach to automation primarily used by individual engineers to solve very specific network challenges. Engineers write scripts for single-use purposes and consequently rarely are designed to be re-usable for similar problems. They tend to be ‘hard-coded’ and are rarely shared, meaning that if 10 engineers need to solve 10 similar problems, it would generate the need for 100 scripts, all doing the same kind of thing!
  • Network configuration and change management (NCCM) tools are an automated means for deploying network changes, keeping a backup and restoring network device configurations, and maintaining an audit trail of configuration changes typically for compliance. They focus on Day-0 and Day-1 and have very limited influence on the in-production network after it has been initially provisioned, especially in the context of what else is happening around the device.

Scripting and NCCM tools enable individual network engineers to speed up their work for identical but repetitive actions, but in doing so propagates the manual approach to all of those processes. Scripting can be considered a brute-force approach which is a tactical response to a single problem. Scripting rarely leverages the knowledge available through other subject matter experts or people that have solved the same kind of problem previously. And because scripts are both generic and hard-coded by design, they are not aware of their surroundings. For example in a hardcoded script, an operator may attempt to set an MTU of “32” for a device even though the valid minimum MTU for any device is 68. That is where human error can become catastrophic to the business. Lastly, scripts can’t be re-used for solving ‘similar’ problems. They must be copied and edited, over and over again. For instance, when applying an updated security parameter across a long list of dissimilar firewall pairs, individual scripts would have to be manually created to address each one uniquely. The concept of “similar” does not exist when writing scripts.08 Campus Branch

The potential benefits of automation extend far beyond scripting and NCCM tools. By expanding automation throughout the entire network, organizations can simplify virtually all facets of network management and in fact looks at network automation as an extension or support mechanism for IT Service Management, since the network is the lifeblood of every IT service.

The next generation of network automation follows a completely different approach from those typically employed by network engineers. It is intelligent and extremely aware of every detail of the topology but abstracts those details so that problem solving can be approached based upon delivered results, not device parameters. It’s a top-down approach, rather than a bottom-up one.

Another important attribute is the portability of knowledge. Intelligent network automation understands the concept of ‘similar’, so as an engineer resolves any issue or task for a particular service, he or she can re-use the same process in other parts of the network in the future, even if the underlying devices are different. Better still, other engineers across the NetOps organization can tap into this library’s past work and leverage it to solve similar challenges with speed and efficiency. It’s like having your problem-solving SMEs’ knowledge available to every engineer, even when the SMEs themselves are not available.

NetBrain applies an adaptive intelligent automation engine to operational challenges in network management.

This includes,

  • Network mapping automation, where a data-driven dynamic map can be created on-demand, effectively replacing the need for static network diagrams. These resulting maps can be consumed individually by engineers, or programmatically within ITSM or NOC systems to document service tickets.
  • Troubleshooting automation, where diagnosis intelligence is first encapsulated through executable Runbooks using no-code methodology, and then available to execute automatically at the moment when a network issue or network attack is happening.
  • Change automation, where comprehensive baselines of delivered network service performance are established, and then background engines continuously test these against the conditions being seen throughout the hybrid network infrastructure in real-time.

Is this a theory? No, not at all. NetBrain PDAS Release 10 and later supports virtually all major network platforms out of the box, allows problem-solving knowledge to be captured without programming, builds a comprehensive digital twin of the entire end-to-end network (including the edge and the cloud and everything in between), and comes with dozens of problem resolutions that can be applied out of the box for the most common errors that occur. And when the network management strategies begin to intersect with traditional ITSM strategies, NetBrain PDAS integrates tightly with all of the major service ticketing systems to extend their reach.