Software's "fail fast" mantra might result in impressive apps and compelling keynotes, but network failures rarely...
win gold stars. And unlike an application bug, a single network error can bring an entire environment crashing to the ground.
When that happens, everyone notices. Yet, even the most dedicated command-line network professional can no longer ignore the call of network virtualization and automation, as legacy architectures struggle to keep pace with rapid-fire application development in the age of DevOps.
Enter NetOps 2.0 -- also known as NetOps, DevNetOps, NetDevOps and Super-NetOps -- a Gartner-coined term that invokes a new vision for the network, along with the evolving suite of tools and tactics for putting that vision into action. While NetOps may refer broadly to network operations of any vintage, NetOps 2.0 represents a paradigm shift from inflexibility to network programmability, with the goal of building agile, dynamic networks to meet the needs of today's enterprise.
Gartner analyst Andrew Lerner called the stakes "substantial," citing network inflexibility as one of the top impediments to critical DevOps initiatives. He and other experts argue network pros -- from managers to executives -- must embrace the ideology of NetOps 2.0 or risk both bottom lines and careers.
Why NetOps 2.0: A history
Evolutionary biologists theorize ancient neural circuitries in the brain drive a phenomenon known as loss aversion, which makes humans strive to prevent potential loss, even at the expense of equivalent possible gain. In other words, we don't like risk or change. Much of the first world now finds the question of immediate survival less pressing -- fewer saber-toothed tigers, more Stop & Shops -- but the neurophysiological negativity bias remains.
Similarly, many of today's network professionals inherited an aversion to risk that stretches back to the dawn of IT, with its origins in the "ancient" internet circuitries of the '80s and '90s. For decades, enterprises issued networking teams a clear mandate: above all else, uptime. In the pre-DevOps environment, availability trumped agility, downtime earned demerits and fortune rarely favored the bold.
Decades of hard, sobering experience taught network engineers to be cautious about change, executing necessary shifts slowly and incrementally and within the confines of real or perceived technical and organizational constraints. This led to complex, ad hoc network designs, with engineers often addressing problems as they arose through manually jury-rigged, snowflake configurations -- each one different and, therefore, each requiring a different fix when broken.
In the short run, managers maintained maximum network availability with minimal disruption. But, in the process, they created an unwieldy infrastructure that proved difficult to scale.
Enter NetOps 2.0: A new approach
Deeply ingrained risk aversion prompted many network professionals to dig in their heels and cling to command-line interface (CLI) when confronted with early DevOps evangelism. And while disruption, innovation and virtualization have radically transformed the systems administration game, many networking teams continue to function much as they always have. Case in point: The vast majority of network changes are still manually configured using CLI, according to Lerner.
Best networking practices now demand a growing level of automation and programmability, as well as meaningful engagement with DevOps initiatives and a greater awareness of business objectives and outcomes -- all part of NetOps 2.0. Rather than resisting change, analysts urge network administrators, managers and architects to accept, manage and use it.
APIs, automation and templates dramatically simplify and speed up provisioning and configuration processes, for example, also taking pressure off of teams to continually re-create the proverbial wheel. Frequent updates and testing can actually engender more stability by improving the integrity of the infrastructure and decreasing its fragilities -- something more or less verboten in legacy network environments, where the cardinal rule is "Don't break anything."
With NetOps 2.0, as attention and energy shift from putting out fires to proactively preventing them, future disasters are averted. In the words of networking blogger Greg Ferro, "Reward firefighting, and you will create a culture of arsonists."
How to implement NetOps 2.0
Organizations need both top-down and bottom-up initiatives to put NetOps 2.0 into practice, with IT leaders working to recalibrate cultural values by normalizing managed risk-taking, rewarding innovative thinking, discouraging blame games after network outages and encouraging challenges to the status quo.
On the ground, networking pros should develop the programming and automation skills necessary to provision, configure and manage API-enabled infrastructure. They also need the soft skills to collaborate effectively with DevOps and DevSecOps teams, both informally and as embedded team members on particular projects.
Lerner suggested team- and skill-building through grassroots initiatives, like automation hackathons and "no-CLI" days. He also recommended adding performance-review metrics that reflect shifting network priorities, such as time to deliver new services, time to detect an outage, and automation as a percentage of total network changes.
While uptime -- the indicator on which networking pros have traditionally been assessed -- remains critical, it takes more than availability to succeed in a NetOps 2.0 world.