Gartner Blog Network


Networking Retrospective

by Andrew Lerner  |  January 2, 2019  |  1 Comment

Around this time of year, many folks take a retrospective look at 2018 and/or make predictions for 2019 (or resolutions). One interesting thing about all these retrospectives is that frankly, networking doesn’t actually change that much in a single year.

While this is not official Gartner position, my theory is that networking only changes about 10% per year. Thus, you could take an entire year off pick things right back up; however take 3 years off and a third of the technology will be new. Simply put, it is hard for most enterprise to rapidly absorb dramatic network changes/innovations, due to a number of reasons (a blog for another day).

We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten – Bill Gates

So this year, I decided to take a look at how networking has changed since 2013 – when I started at Gartner and roughly the lifecycle of many network devices.  Looking back, I found 5 major changes compared to 6 years ago, in no particular order (Spoiler alert:  SDN isn’t one of them, and neither is intent).

VMware – VMware is now a major networking vendor. Six years ago, they had vSwitches and recently purchased Nicira, but everyone was essentially waiting to see what they were going to do with it.  Fast forward to today, NSX comes up in most client data center networking discussions and VeloCloud (owned by VMware) comes up in most WAN conversations.  Which brings us to big change #2…

SDWAN – SDWAN wasn’t a thing 6 years ago, and now is the defacto discussion when organizations refresh routers or migrate to a hybrid WAN topology.  It has gone from early days (dozens of customers) to a mainstream (thousands of customers) product in just a few years (yes – I realize this contradicts my earlier statements about network change…also another blog for another day).  The reason is because there is/was sooo much pent up demand for better WAN products.  For context, the hottest WAN topic 6 years ago was Wan Optimization, which has moved from dedicated market to an embedded feature.

Light Weight Load Balancers – Back in 2013, we got a ton of calls about Application Delivery Controllers, which were deployed primarily as hardware and front-ending hundreds of apps in a given environment, managed by IT and often the networking team.  Fast forward to today, the market is much more fragmented.  Companies are refreshing their traditional ADCs sure, but many new cloud-native apps aren’t behind traditional ADCs.  This functionality is increasingly implemented on a per app basis and by lighter weight and lower cost alternatives, and being led by Cloud, DevOps and even line-of-business teams.  Thus, there has been a massive shift towards public cloud providers’ embedded load balancers, and open source options like NGINX and HAproxy. Moving forward, service meshes will consume even more of that functionality within microservices environments.

Network Automation – Back in 2013, network automation was a nice to have, and mostly aspirational.  Nearly 90% of networking changes were manually-driven. Now, automation is viewed as a foundationally important, and the way that networks should be run moving forward.  That said, we still have a ways to go before the CLI is dead

Microsegmentation – While I personally don’t love the term, the reality is that microsegmentation is now a huge part of many conversations in networking – particularly in the data center.  There simply wasn’t any awareness back in 2013.

Happy New Year…

Regards, Andrew

Additional Resources

Category: adc  devops  networking  sd-wan  security  wan  woc  

Tags: cli  haproxy  microsegmentation  nginx  sd-wan  service-mesh  vmware  

Andrew Lerner
Research Vice President
6+ years at Gartner
21 years IT Industry

Andrew Lerner is a Vice President in Gartner Research. He covers enterprise networking, including data center, campus and WAN with a focus on emerging technologies (SDN, SD-WAN, and Intent-based networking). Read Full Bio


Thoughts on Networking Retrospective


  1. ahenning says:

    CLI would live until AI replace humans as primary coders, because CLI is just a high level programming language. Any tool used to replace what is done on the CLI is just another form of high/low level code. The more things change, the more they stay the same.



Leave a Reply

Your email address will not be published. Required fields are marked *

Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.