Blog post

Checking in on the Death of the CLI

By Andrew Lerner | January 04, 2018 | 1 Comment

NetworkingJust Published

Every year around this time, a ton of pundits make predictions. Some of them are good, some funny, while others… not so much. I’m not a huge predictions guy, but recognize there is solid value in making them. Gartner does publish official predictions across multiple technology areas (including networking) every year.

However, the point today is to check-in on a previous prediction, because it seems that very few people actually do that (Kudos to those of you that do). About 14 months ago, we predicted that By 2020, only 30% of network operations teams will use the command line interface (CLI) as their primary interface, down from 85% at YE16.  So basically after a ~20 year run as the primary interface, we were predicting a pretty dramatic drop-off in 37 months.

So, to check-in… The survey results below are based on audience polling from a recent Gartner conference, so note it is a biased and small sample size, but this does match what we’re hearing from clients. Things have improved versus last year (or maybe just gotten “less bad”) when CLI usage was easily in the 80-90% range.


However, the real kicker is that when we asked folks where they plan to make the most strategic networking investments in 2018, “network automation” won by a landslide with 49% (n=129).  Other options included cloud-managed LANs, Ethernet Fabrics, Service Provider On-Demand services, Intent-based networking, SDWAN/vCPE, SDN/Overlays and NFV (basically all the hot buzzwords) but none of these received more than 16% of the votes. This is a very good thing, because while we still have a long way to go as an industry, this indicates the mainstream is finally starting to prioritize network automation. I believe we can get still get to the point where the CLI is not the primary interface within 23 months (so you’re saying there’s a chance).


This would be great for our industry because (among other benefits), automation leads to faster delivery of services, scales better with growth, reduces outages, and allows folks to move away from tedious lower-level tasks.

Happy New Year and Happy Automation,

Regards, Andrew

PS – We do have a published Market Guide for Network Automation (paywall)

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Leave a Comment

1 Comment

  • David STEIN says:

    While I think automation MIGHT be a good thing, I do not think it should replace the CLI. It should complement it in certain cases but definately not replace it.

    The problem in capitalism in general and in the IT in particular, is that it’s always about replacing something that work with something that is always marketed to be better.

    Let’s take the smartphone industry as an example… The new smartphone coming out is always A REVOLUTION compared to the previous one that came out only 12 month before.

    About automation, truth is, nowadays not a lot of companies are increasing their OPEX when they have any. And we all know that it DOES take a bit of training to go from CLI to python, JSON, ansible…etc… If companies don’t want to handle the cost of the training, and I am sure that most of them won’t, then who will ? And in this very likely case, what is that chance for the CLI to take over ?

    Also, we have to stop letting buzzes dictat us what we should do. The CLI is the best way to understand concepts, how to implement and troubleshoot. Packet captures are even better for that matter… But the more we will overlay (with things like automation) the less we will understand how things work and thus how to troubleshoot them.

    I am extremely doubtful about the ability of automation to troubleshoot any potential issues (such as IOS bugs for instance). That’s why I think automation should only be a complementary tool.