Only 37% of Enterprises are Fully Successful with Network Automation
by Shamus McGillicuddy, Research Director, Network Management
Over the last several years, Enterprise Management Associates (EMA) research found that enterprises are actively expanding their use of network automation tools. For instance, EMA’s “Network Management Megatrends 2018” research found that 92% of network managers have a formal initiative to expand their use of network automation, and 70% called that initiative a high priority.
To understand these automation projects more fully, EMA has completed a new study, titled “Enterprise Network Automation for 2020 and Beyond.” The research is based on a survey of 250 IT professionals directly involved in a formal network automation initiative and phone interviews with a half-dozen practitioners. EMA found that only 37% of enterprises are fully successful with these automation efforts. Another 46% claimed they were somewhat successful, with room for improvement. Thirteen percent said it was too soon to tell whether they were successful or not, and 5% claimed they had been unsuccessful.
EMA asked survey respondents to identify what they find difficult about network automation, from an automation product perspective and a business perspective. One emerging issue appears to be dollars and cents.
Network Automation Tool Challenges
When EMA asked research participants to identify their biggest challenges with the network automation tools they have worked with or evaluated, the top response was price. The tools are simply too expensive.
The top secondary challenges were difficulty of implementation, scalability, and breadth of automation, because there are some network device features that cannot be automated. Thus, enterprises are finding that it’s difficult to get network automation tools up and running, whether those tools are commercial products or homegrown solutions. Scalability is a common problem with network management tools in general, since software development is strained by the volume of data it must process and the number of commands it must push out. Breadth of automation suggests that one tool can’t get the whole job done. This latter finding may explain why 91% of the enterprises in EMA’s survey use more than one network automation tool.
Network Automation Business Challenges
EMA also asked enterprises to identify the business challenges they encounter with their network automation initiatives. The top response was security risk. For one thing, enterprises may be concerned that their automation tools are vulnerable to exploitation. More likely, they are worried that automation might introduce unexpected vulnerabilities into the network. For instance, a poorly implemented tool might allow external entities to access a secure database.
The second-biggest business challenge is budget. Enterprises lack the money to purchase tools or hire people to build them internally. This echoes the top product complaint mentioned earlier—price. Thus, not only are automation tools expensive, but enterprises are struggling to get budget allocated.
Enterprises are also struggling with a cultural resistance to network automation, which is the third leading business challenge. Network professionals simply don’t trust the automation that has been implemented or they think the old way of managing the network is more effective. In other cases, network professionals struggle to understand the full value of automation technology. As one research participant, a network reliability engineer with a midsized global media and entertainment enterprise, told EMA, “Even the people who are enthused about automation struggled to get on the right page with it. People loved the config capabilities, but they spent so much time manually collecting information from the network for troubleshooting. They struggled to identify how automation could affect them in that space. They couldn’t visualize how solutions could help them.”
“There is a fear that you would bring the network down [with bad automation],” said a network engineer with a very large global pharmaceutical company. “At one point, I pushed a change that took out one line of code, and it caused a CPU spike on 50 devices on the network. It was not pretty.”
Network Automation Skills Gaps
Finally, 96% of research participants revealed to EMA that they are working to close skills gaps with network automation. The top priority is training network managers on how to use network automation tools. However, there are several other less remedial training issues that must also be resolved.
The second-highest skills gap issue is cross-team expertise. Enterprises find that automation requires their network professionals to learn how other parts of the IT organization work. They need to understand processes and management tools.
Next, network automation often involves some advanced analytics or AIOps technology. The third-biggest skills gap is data science, or data engineering. Enterprises need people who know how advanced algorithms work and can support them.
The fourth most pressing skills gap is with new network infrastructure devices. Network automation often leads enterprises to adopt new network devices because legacy equipment might lack APIs or other interfaces for automation. In fact, 45% of the enterprises in this study said that network automation was a leading factor in their adoption of new networking vendors.
EMA’s new network automation research looks at more than just the challenges that enterprises encounter with network automation. The full report explores the technologies that enterprises are using, the management processes they are automating, and the devices they are targeting with automation.