The Last Router

The year was 2012 and a new reality was setting in for the server market. It had been long dominated by companies like HP, Dell and IBM – who together had commanded a whopping 80% of the global server market and hence were Intel’s main customers for server CPUs. But in 2012 things started to shift dramatically and very quickly. In 2012, Google had suddenly become Intel’s fourth largest CPU customer. The Cloud had started to cause a tectonic shift in computing.

Fast-forward to 2020 and read how dramatic that shift has been in this article in Network World. Let me quote some key data:

• “ .. Enterprise tech crossed a significant line as the decade ended. For the first time, enterprises spent more annually on cloud infrastructure services than on data-center hardware and software, according to Synergy Research Group ..”
• “ .. total spending on cloud infrastructure services in 2019 will reach $97 billion, a 38% increase over the prior year. Ten years ago, that spending was near zero ..“
• “.. From 2009 to 2019, average annual spending growth for cloud infrastructure services was 56%, according to Synergy, while on-prem hardware sales grew only 4% on average ..”

No, the server market hasn’t disappeared, and companies like HP and Dell are still relevant. But their market share in the server market has declined into the 15% (each) range, trending further down in the future (IBM abandoned the server market in 2014). The big winners are IaaS services from the big cloud providers, clearly – and to a lesser degree, white boxes (which big cloud providers mostly build themselves).

The lesson is that, in the digital era, enterprises prefer to invest into technology that immediately touches and benefits the business layer, which is where technology truly mints $revenue$. Technology needs to directly support differentiation and drive expected business outcomes. In the computing world, clearly that means that agility in application innovation is the priority and dealing with hardware and OS installs and updates is a hindrance rather than an enabler. Hence the shift to the cloud, containers, orchestration and even serverless computing.

And where the computing world goes, so follows networking, albeit with some years-long delay. Famously, Prof. Scott Shenker from Berkley University, one of the pioneers in Software-Defined Networking, wondered in a presentation at Ericsson Labs whether it was because application software designers are much smarter than network architects (don’t worry, he was joking!).

And several leading analysts’ qualitative and quantitative predictions for the SD-WAN market bear that out (I mean the shift towards a Network-as-a-Service or Managed Services model, not the fact software architects are smarter than network architects :-D). Let’s look at some:

• Gartner: “.. SD-WAN managed services will grow at a CAGR (Compound Annual Growth Rate) of 84.7% between now and 2022 ..” (Note: In the same research note, the SD-WAN hardware equipment market’s CAGR is 30%.)
• Nemertes: “.. One factor accelerating adoption is the rapid profusion of managed solutions .. 30.1% use a managed SD-WAN service.. This represents a huge jump in SD-WAN service adoption in a short time ..”
• MarketsAndMarkets: “.. Plenty of channel experts also have suggested that customers prefer SD-WAN as a managed service. Customers lean on the side of having more help .. they need to be able to have the peace of mind that someone is managing and monitoring that network to ensure it’s operating at its prime ..”

We all know that the wide-area networking market is still primarily a Do-It-Yourself environment, where network engineers take manual, loving care of their “router pets” (the analogy in explained in this previous blog). And once they’re done with the installation of their new routers -even in the brave new SD-WAN world- they’ll spend over 70% of their time keeping-the-lights-on and trouble-shooting, with barely any time to think about how to unlock new business benefits from the network infrastructure. This operational overhead costs enterprises worldwide a stunning $60B every year according to a much-cited McKinsey study.

But think about the primary business benefits enterprises should focus on to drive value from their network in the digital era:

• Employee productivity fueled by superior business-relevant application performance.
• A constantly optimized topology to support the dynamic nature inherent to cloud adoption.
• Real-time, consolidated end-to-end network and application performance visibility.
• The ability to turn up/down sites, services, capacity anywhere within 24-48 hours (and not weeks or even months).

Real-world deployments indicate that Do-It-Yourself SD-WAN is still hampered with many of the pitfalls and complexity of traditional networking, showing that we are probably moving too slowly in the networking world when it comes to addressing enterprises’ needs for agility and speed.

There is no doubt in my mind that the answer to that is a Network-as-a-Service model. I know you’ll be thinking “How conveniently self-serving, Paul, you are after all working for Aryaka!”. I can’t deny the obvious: I want Aryaka to continue to grow fast and continue to succeed as a visionary in the SD-WAN market. On the other hand, I think you have to admit that it simply makes sense that the networking world ultimately follows the design patterns that have taken the computing and application world by storm – and that means accelerated, wide-spread as-a-Service adoption, as evidenced by all the market stats cited previously.

Furthermore, if you read the blog I co-wrote with my friend Klaus, you will see that we readily admit we -as Aryaka- are still growing our capabilities to address even more enterprise use cases.

The future of networking undoubtedly lies in abstracting complexity through intent and orchestration. In providing enterprises choice in multi-cloud options and security to truly be able to dynamically tailor their network infrastructure to their ever-changing business needs.

Network engineers need to become -like their DevOps peers in the application world- agents of constant change and improvement. Think about it: How can an enterprise network with static policies ever attempt to optimally support a container-based app which Kubernetes just spun up over several geographically distributed pods to meet demand? Network needs have changed dramatically and it’s hard for us network engineers to realize how fast that has happened. We are still collectively trying to linearly extrapolate in a time of disruptive, exponential change. The DIY SD-WAN ultimately is still a statically configured network and cannot address the needs of the Cloud-First era.

The Aryaka approach foreshadows the inevitable future of enterprise networking, overcoming the challenges of a fractured middle-mile, the static network topology constraints that no longer suffice in a Cloud-First world, the undeniable need for a multi-vendor and open approach to security and several more.

Just like in the server world, in the networking world the relevancy of the router -even SD-WAN enabled- is going to fade. Don’t be the last network manager that buys a router a few years down the line, explore the unstoppable trend to adopt NaaS.

Check us out and see a demo!