Good engineers will tell you that the best way to solve complex problems is to break them down into smaller, more manageable pieces. Therefore, it should come as no surprise that enterprise IT has organically evolved to become a collection of individual silos, each aimed at solving a part of the larger IT problem. Though functional as a whole, enterprise IT more closely resembles a mosaic of technologies, tooling, processes and, of course, the people expected to manage the pieces.
But what happens when the natural evolution of IT requires that these pieces come together? Where will convergence happen first? And what must IT leaders do in preparation for this inevitability?
Multicloud as an Operational Driver
There are many reasons for enterprises to break down the silos that exist within their IT teams. Cost and efficiency are at the top of virtually every CIO’s wish list. Unfortunately, without a compelling reason to force the change, there is typically too much inertia to change course and too much risk to justify the effort.
But there are industry-wide movements afoot that will likely upend traditional IT, either directly as new architectures are adopted or indirectly as competitive advantage forces strategic changes at a company level.
Multicloud is one of these tectonic shifts. The idea that workloads will move to the cloud is disruptive enough, but multicloud adds an operational element as enterprises look to exploit diverse underlying infrastructure while converging operations to help drive more agile operations and contain costs that come with sprawling infrastructure. Here, it is important to note that multiple clouds and multicloud are two different concepts. Whereas companies are already managing multiple clouds—a workload in AWS, a SaaS offering and on-premises workloads—the shift to multicloud is only just now emerging.
More simply, in a multicloud world, the central thesis is not that workloads will move dynamically across different clouds, as applications are not that portable and the complexity of making them so would outweigh the cost benefits. Rather, multicloud is about finding consistent ways to express application policy and control regardless of where the application is hosted. It’s about converging on a narrow set of operations so that experience can be uniform even when infrastructure is not. Ultimately, policy and control will have to extend all the way across the WAN, encompassing elements like the data center edge or the WAN gateways that connect to it.
There will be an operational transformation that takes place amidst the larger move to cloud and that transformation will ultimately have implications on how traditional IT silos are treated.
It’s Always the Wi-Fi
A familiar refrain in almost any group that has support responsibility for enterprise infrastructure is that it’s always the Wi-Fi. Of course, this isn’t true, but every user that experiences an issue has limited visibility beyond how applications and services are accessed. So, whatever the problem is, the first thing blamed is almost always the Wi-Fi, putting the support team in a defensive position.
The Wi-Fi must be proven innocent, before the rest of the people required to appropriately diagnose the true problem can be activated.
The Power of Unified Operations
Imagine that there is a configuration error on the WAN gateway that introduces an MTU mismatch across the WAN link which can result in packet loss over the WAN for some subset of traffic. Because the issue is not persistent, it can be difficult to diagnose.
In this case, the wireless management system can spot intermittent packet drops. A health check of the wireless LAN might reveal that the issue likely resides elsewhere. This guides teams to examine the next hop in the chain. Understanding what changes have been made recently helps narrow down the issue to a configuration error on the WAN gateway.
This example is simple but illustrates of the power of unifying operations. If a troubleshooting workflow spans boundaries, the only operational path forward is to either hand off between teams or to unify the model.
Moving Beyond the CLI
In some ways, there is a natural dampener on any progress in the pursuit of unified operations. Enterprise IT is dominated by the command line and despite decades of desperately pursuing a more automated environment, the vast majority of IT operations is still conducted through the command-line interface (CLI).
Coincident with the multicloud movement, though, has been the transition to software-defined everything. By introducing abstracted control, architects and operators have begun the transition from command-by-command to intent-based control. This provides natural convergence points for operations. It starts with expanding visibility across boundaries, but this will quickly lead to workflows that can be executed across what used to be isolated silos.
Enterprises that have aggressively pursued software-defined ambitions alongside their automation agenda will find that they are set up well for the operational transformations that accompany cloud and multicloud.
Applying Converged Operations
The natural next step in this journey is to begin purposefully pursuing workflows. While much of the automation discussion starts in the data center, it might make more sense for enterprises to look at the campus and branch, unifying the wired and wireless LAN first, followed shortly by the LAN and WAN.
The move to SD-WAN provides yet another opportunity to bring the software-defined WAN together with the software-defined campus and branch. Indeed, 2020 will likely see an industry-wide expansion of SD-WAN to SD-enterprise. Of course, engineering teams need not wait for the marketing phrases before they begin.
Just the Beginning
Ultimately, this is all just the beginning. The software-defined movement might accelerate converged operations, but it too will be the end result. If automation is central, it would seem that “AI-driven” will be the next natural evolution, allowing the transition of manual cross-domain operations to something more proactive and eventually predictive. Thinking now about analytics, telemetry, automation and AI will help accelerate the next wave of change. After all, the only constant in today’s enterprise IT is precisely that: change.