At NetBox Labs, we frequently hear from IT leaders and network engineers on the progress of their IT and network automation initiatives. Network automation has lagged compared to automation in other areas of IT. We often hear that “keep the lights on” work prevents network engineers and architects from progressing their network automation strategies. Networking teams are underwater, and research from EMA in 2022 reported that only 27% of networking teams report “success” in their day to day jobs of managing complex network infrastructure, down from ~50% 5y ago. We see and hear about the impact of network complexity and manual network operations every day.
Last year, NetBox Labs began investing directly to understand and solve the problems preventing teams from progressing their network automation strategies. And we invested not just in our products, but in our community. We did this through initiatives like NetDevOps Days, which brought practitioners together to share success stories and learnings, through sharing takeaways from teams like Dartmouth College whose network automation strategies have demonstrated big success and impact, and even by working with practitioners across our community to develop a common reference architecture for modern network automation.
As we learned with our community, we decided it was important to contextualize the conversation around what’s holding back network automation with real data. We decided to partner with Shamus McGillicuddy at EMA to better quantify the current state of network automation, learn more about how teams are making progress, and find out what’s holding them back.
We sought to answer questions like:
- Are teams as behind as they perceive themselves to be?
- How many teams have executed their network automation strategies to some degree of success?
- Are there common practices among successful companies that have made it to the later stages of their network automation journey?
- Is there a disconnect between the latest technologies and their actual adoption at scale?
In this post, I’ll discuss what I found were some of the biggest industry trends uncovered by this research with EMA, and what they mean for the network automation community. If you’re interested in diving deeper into the research, download the research report here or listen to our recent webinar with EMA’s Shamus McGillicuddy who led the research design, implementation, and analysis.
Network Automation: Trends and Adoption Drivers
Cloud, security, and network complexities are driving network automation spend
A few weeks back I met with the networking and datacenter teams at a major media company in New York that depends on NetBox in their infrastructure operations. I found it fascinating to hear that a huge chunk of their energy was devoted to information sharing with their security team, which had a constant stream of requests for data about the network, IT infrastructure, changes, anomalies, and more. Of course, at NetBox Labs we’re investing in making network data accessible to security, cloud, finance, and other teams through the tools they use natively like ServiceNow or Splunk. But this conversation illustrated a huge trend that’s driving networking teams to automate: security and compliance needs are creating operational drag and introducing network complexities, and that’s leading networking teams to automate.
Similarly, we’ve heard from many networking teams that the adoption of cloud or hybrid cloud architectures by their applications teams is setting expectations for the enterprise infrastructure. “Customers” of networking and IT operations teams expect self-service, dynamic networking setups and are conditioned by their cloud experiences. Over and over we hear this is driving a need for network automation.
And this is exactly what Shamus found in his research: security and cloud are consistently among the largest trends driving network automation in enterprises closely matching what we’ve seen anecdotally among teams adopting NetBox to accelerate their automation journeys.
Network automation success requires adequate funding
Going into this research, I wouldn’t have been shocked to learn that network automation projects with larger budgets would report more successful outcomes. But I was still surprised to see how clearly this story plays out in the data.
While 80% of organizations that claimed “complete success” with their network automation reported their project had been fully funded. Organizations that said they were only partially successful were only fully funded 57% of the time, and those whose projects partially failed had only received full funding 29% of the time. EMA also found that 76% of organizations expect their network automation budgets to grow over the next two years, but fewer – only 21.2% – described that growth as significant.
In other words, network automation spend is rising, but it might not be enough to guarantee desired outcomes. But this research is starting to quantify the correlation between adequate funding and success in network automation projects. If you ever needed the data to back up your budget requests, this is it.
What defines a completed network automation project?
At NetBox Labs, we’ve learned that implementing network automation is a journey, not a step function, so it was interesting for us to understand how the industry views their progression down the network automation path.
To our initial surprise, the research found that 28% of respondents feel they have “completed” their network automation projects. But peeking behind the curtain, it’s apparent that “completed network automation project” isn’t the same as “automated network”.
The survey mostly polled senior executives at larger enterprises, not practitioner network engineers in the network management trenches. In fact, EMA’s research found that even in organizations with completed projects, about 57% of all network tasks are still manual, meaning only 43% are automated today. But that’s changing – respondents think that in just two years, the scales will tip to automation: 66% of network management tasks will be automated by 2026.
In short, this finding speaks to how organizations feel about the level of network automation that is required for their current needs, and the overall success of their efforts. Even though early network automation projects have been completed, they haven’t been particularly ambitious, and there are still plenty of tasks awaiting automation. There’s lots more to be done.
And just as with network automation overall, there’s lots more to learn from the findings in Shamus’s study. In our next post of this three part series, we’ll dive deeper into what practices drive successful network automation. Stay tuned for more to come!
[End of Post 1]
