Network automation tools improve agility and efficiency, lower costs and reduce errors. Network automation tools have been around for 20+ years, but there is renewed interest with concepts like GitOps and NetDevOps gaining popularity. We just published a Market Guide for Network Automation Tools (paywall) and here are a few takeaways.
The market is extremely fragmented, as there are 50+ viable network automation tools adopted in enterprises and at least 15 of these tools that have reached an installed base of over 1,000 enterprises. However, despite the abundance of tools, ClickOPS is still prominent as less than 35% of enterprise network activities are automated today.
While the overall level of automation is just mediocre (at best), there are a small percentage of enterprises (and I’m not talking about AWS, Facebook and Google types) that automate more than half of their network activities. Thus, it is kinda like a HAVES vs HAVE-NOTS thing going on. Some of the HAVES even restrict manual CLI access to devices. Further, these “HAVES” are quite vocal in the broader community and tend to create a false sense of widespread network automation in the industry (and vendors tend to cater to them).
In the report we go into the barriers to network automation. It is interesting to note that despite “conventional wisdom” that network folks are worried about “automating themselves out of a job,” this is not something Gartner actually observes. The report includes a bunch of pragmatic and specific recommendations, including to start with small network initiatives and iterate by focusing on nonchange and/or nonproduction activities. We recommend automating “quick win” activities such as troubleshooting, baselining and archival. Specific examples of “quick wins” include
- Autopopulating trouble tickets with network information.
- Creating an automated network availability baseline, such as verification of uptime for specific services/applications.
- Creating an automated network performance baseline, such as an automated test that verifies and records whether latency to specific services/applications is within a threshold.
- Automating archival of device configurations.
- Automating archival of network operational state (device routing/switching tables, for example).
- Automatically enabling/disabling network monitoring tools during a change.
A few other snippets from the report: We estimate there are over 30,000 enterprises using open source Ansible alone, which is popular due to its agentless nature along with support for multiple devices and widespread support from commercial networking vendors. We predict that dedicated network change and configuration management (NCCM) products will be increasingly delivered as a feature in broader automation and infrastructure platforms (vs a standalone product) — and available via an “as a service” model.