The Network Automation Paradox: Why People are Key to Its Success
The saying “doing the same thing over and over again and expecting a different result is the definition of insanity” could very well be used to talk about network automation.
For the better part of two decades, network professionals—vendors and customers alike—have pursued their automation ambitions. Entire companies have been born (and died) chasing the dream. Tens of thousands of collective staff-hours have been poured into an effort that, by any meaningful measure, has largely been a failure.
At what point does the industry accept defeat and try a different approach?
Solving for adoption
Production-side problems in networking exist because no product, solution, or service meets the need of customers. Essentially, the need will go unmet until someone produces a solution. In a high-tech arena like networking, this typically means that either the problem is not yet solved (the technology does not yet exist) or that it is overstated (no one wants to build it).
On the other hand, consumption-side issues in networking occur when despite having a solution, users are incapable of using them or do not want to use them. Again, in the high-tech space, this typically happens when the cure is worse than the disease—either because the solution is too costly or perhaps because it is prohibitively difficult to use.
In networking, the automation adoption problem is largely a consumption-side issue. The tools have existed for decades to do things from scripted workflows to centralized control. Despite those tools, though, the vast majority of network operations are still manual, driven by fingers on keyboards behind some device CLI.
If automation represents a consumption-side problem, then the key to moving the industry forward is in fostering more adoption. This means the industry needs to develop a working theory on why adoption has lagged despite bold pronouncements that everything should be automated.
Start with why
In his famous book, Simon Sinek encourages leaders to “Start with Why.” If people do not understand why they should do something, they are frequently not inspired to act.
Similarly, when it comes to network automation, the industry has never really correctly identified why automation matters and doesn’t understand the “why” behind it. For too many companies, automation is about driving down cost by eliminating the people required to make networks work. For others, it is about speeding up operations. But neither of these capture the real objective of network automation.
Why not cost reduction as a motivator?
There are several problems with cost reduction as the primary driver behind automation, including that it relies on a workforce that will ostensibly be relieved of their jobs to carry out the cost change faithfully. While people support their companies, at the end of the day, they are only human. And asking an entire population of network engineers to hasten their own demise seems silly. At best, it’s not an ideal way to drive change.
Beyond that, the cost objective is actually misplaced. Companies that initiate an automation agenda in pursuit of cost-cutting will find that their OpEx will go up before it goes down, as it requires keeping the lights on while developing a new practice. While there are efficiency advantages, the goal is to drive scalable growth, not accelerate the cutting of costs. Growth is the agenda while cost is the inhibitor – and that takes automation from a CFO initiative to a technological imperative.
Why not speed as the end goal?
If not cost, automation is surely about moving faster, right? Except network operators can already move as fast as they want. There is nothing inherent in networking that forces things to move slowly. If the objective is speed, the industry’s reliance on ITIL and waterfall process could simply be set aside, allowing teams to make ad hoc changes at whatever rate they desire. Something as simple as scripting can even move things faster.
But speed isn’t actually the problem.
If not cost or speed, then what is the automation imperative?
The online parody account Devops_borat once tweeted: To make error is human. To propagate to all server in automatic way is #devops.
While a funny characterization of DevOps, there is some wisdom in this line. The thing that prevents network teams from moving fast is the fear of breaking things. Networks are notoriously fragile. In fact, ask a network team to describe their network. Where does the descriptor reliable show up in the list? It likely doesn’t.
The goal of automation is to improve reliability. It is about systematizing how things are done so that success is a predetermined outcome. If a network operator were certain his or her changes would work, he or she would move more quickly, and in doing so, become more efficient.
It’s not that cost and speed don’t matter. Rather, they are the byproducts of automation, not the foundational reasons why it matters.
Back to people
This brings the conversation back to people. While the industry talks a lot about this product or that tool or some coding project, the real foundation of automation is people.
A product doesn’t exist that can provide enough contextual awareness to eliminate the need for knowledge. There isn’t a tool that can replace the complex understanding required to design and operate the network—a system of systems connecting other systems. But there are methods, supported by products, tools, and projects, that can help make the practice of networking more reliable. The key to unlocking those methods starts with understanding the problem space, and it ends with an industry-wide push for greater adoption by focusing on the people behind it.
For some, a change in operations represents risk. Individuals and companies that see their livelihoods in established methods will move only begrudgingly. This puts a greater responsibility on the collective to carry everyone forward.
Ultimately, it’s not that 20 years of desire have been wrong. Network automation is indeed a worthy endeavor. But the industry’s inability to move forward despite that desire means it’s time to try a new approach that focuses on people in order to accelerate the pace of automation adoption in the industry.
Recommended For You
Low-Power WANs offer an alternative to 5G for connecting a fast-growing array of basic devices and sensors that transmit small amounts of data.
An effective network visibility strategy requires understanding the technical, financial, political, and legal aspects impacting your network operations.
Emerging organizational structures for IT include placement of IT pros in user areas and departments forming their own "micro IT's."
Comparing a good and bad trace helps identify performance issues. Dynamic baselining can be used when you do not have a good trace to reference.
Combining commodity server platforms and FPGA-based SmartNICs will allow network applications to operate at hundreds of gigabits of throughput with support for millions of simultaneous flows.
SD-WAN implementations are on the rise thanks to the potential cost savings, increased network resiliency, and better application performance they deliver.