The vast majority of enterprise IT organizations that pursue a network automation initiative will encounter a skills gap, but the most challenging gap to close will not be programming know-how.
Enterprise Management Associates (EMA) recently published "Enterprise Network Automation for 2020 and Beyond," a research report based on a survey of 250 enterprises engaged in a formal network automation initiative. The research also included one-on-one phone interviews with six practitioners of network automation.
The research found only 3% of enterprises had sufficient network automation skills and knowledge on their network team to support an automation strategy. Nearly 40% said they devoted significant time and resources to training their personnel.
What automation skills should network engineers have?
When EMA asked enterprises to identify the skills and knowledge they most needed to acquire for network automation, scripting, programming and software development skills were at the bottom of the list.
Instead, 29% of survey participants said they needed to train their personnel on how to use the network automation tools they had installed. These tools might include homegrown tools that an internal software development team built. In those cases, the enterprise would have to put together a custom training program. More likely, these organizations were using one or more commercial automation tools. Vendors of commercial tools usually offer training to customers.
The second most pressing priority, as selected by 24% of organizations, was cross-team training on processes and tools in other IT domains. In other words, once network automation is in place, network engineers and architects will be spending less time performing command-line interface ninjutsu on box-by-box network management. Instead, by working as part of a cross-domain services team, they will need to collaborate with and understand how other parts of the IT organization work.
The No. 3 priority for network automation training was data science knowledge, as indicated by 22% of organizations. Enterprises need people who know how to work with advanced algorithms, data structures and data models. This finding indicates that advanced analytics and AIOps are a critical component of a network automation initiative. In fact, 85% of survey participants said advanced analytics was a component of their automation strategy.
The fourth-highest priority in network automation skills is training on new network infrastructure devices, according to the EMA study. This last finding is not surprising, since 89% of enterprises said network automation had been a contributing or primary driver of their organization's decision to adopt a new network infrastructure vendor somewhere in their network.
However, don't expect widespread ripping and replacing of network hardware as part of an automation initiative. But some new network hardware platforms are more automation-friendly than previous generations of infrastructure products, both from market leaders and startups. In some parts of the network, an enterprise may adopt a new platform in the interest of advancing automation.
Coding and scripting still important
While software development, at 16%, and scripting, at 14%, were the lowest training priorities, these network automation skills are in demand. After all, 54% of these enterprises said one or more of their automation tools require coding and scripting to maintain their value. Thus, IT organizations need someone on the team to get this done.
As a network reliability engineer with a midsize, global media and entertainment company told EMA, "I was hired for automation. I had software development experience with different automation tools. I also had network engineering skills."
In some cases, enterprises have plenty of application developers in-house already, but those developers lack networking knowledge. This fact is reflected by this account from a network automation specialist with a large software company: "[Management] took me, a networking person, into a [network automation] programming role because there were visible gaps within the programming group where they didn't understand networking."
While each enterprise will vary in where exactly it finds skills gaps, one thing is clear: Network automating will require training of personnel. Some of that training may be easy to find. Other skills may require time and patience to develop. Regardless, IT leadership needs to recognize this gap and plan accordingly.