How SD-WAN will make the cloud much, much bigger

The user-defined WAN will give greater access to corporate resources, but also require more cloud support and faster links to the cloud.

Bill Baker Jul 24th 2018

Though I no longer actively participate in it as a pioneering player in the networking space I have always kept a watchful eye on the market and I am seeing yet another disruptive force known as SD-WAN (Software-Defined Wide-Area Networking) finally gaining momentum.

For starters, SD-WAN is an extension of Software-Defined Networking (SDN). As the term implies, SDN aims to automate (virtualize) various network functions that are currently touch-heavy.

Network architects talk about separating the control plane from the data plane ad nauseum but that is just the starting point. The ability to virtualize numerous network functions from a central location and thus create an abstraction layer in a manner that is custom-tailored for each enterprise – and, by extension, perhaps for each user - has been the Holy Grail of networking for years if not decades. Tom Nolle, president of CIMI Corporation and a longtime proponent of all things SDN, is spot on when he states that: "SD-WAN is absolutely critical, because it is the vehicle most likely to bring true virtualization to networking. Without virtualization in the network, virtualization in the cloud or data center will never be fully realized or effective."

The need to be connected to the mother ship is what brings the Cloud into its meaningful existence because we live and work at the edges of the Cloud. SD-WAN is not just a market but a platform as well that will eventually evolve into user-defined WAN (UD-WAN). To clarify, the term applies to enterprise users and not consumers. And the purpose of SD-WAN is to connect and fully integrate the very edges of the enterprise – be it corporate headquarters, branch/remote offices or the mobile millions. In other words, us, the users.

But if we look at the concept of the cloud it is pretty clear that it is referenced in an abstract form. After all what is this cloud thing? Some physical space in a non-descript windowless warehouse? Without its tentacles, the cloud is nothing more than a collection of computers, storage and cooling systems created by geeks and for what purpose? It is those very tentacles in the form of wide-area networks (WAN) that give the Cloud its purpose. And given the explosive adoption of cloud-based applications (Box, Dropbox, Salesforce, SAP, Slack, etc.) cloud computing is not a fad, it is here to stay. However, that is just the beginning.

The migration towards digital transformation has already started with countless enterprises globally either evaluating and/or transitioning towards this new and very promising paradigm. To use an analogy, a bullet train cannot run on top of railroad tracks laid over a century ago. New and innovative methods are needed to fully leverage these emerging and very promising models. Enterprises can see the immediate benefits of the cloud, such as lower cost of ownership, reduced bandwidth expenditure via hybrid networks, and greater control of security and network-policy administration.

To get a full appreciations of the potential of UD-WAN and its effect on cloud-based computing we need to examine what happened to the mobile world once Apple launched its App Store (and iTunes) and the subsequent effect they had in the consumer sector. With each new app and new song release the App Store/Google Play started to expand like the mobile versions of the Big Bang. According to Statista, the number of downloads in 2017 was 178 billion, expected to grow to 258 billion by 2022. That is a huge number even by Dr. Carl Sagan’s standards. I’m not suggesting UD-WAN will generate those numbers, but it will certainly generate a lot more than the most current analysts’ optimistic projections. No one predicted the numbers the App Store/Google Play would generate.

To put things in their proper perspective let’s also examine how the edge operates. The edge is a physical location (branch office, remote office) with employees connected via high-speed connection to HQ/enterprise. In addition to voice calls there are numerous apps that enable the Edge to integrate with HQ’s cloud-based operations. For the most part these employees are not stationary; they move about staying connected via mobile phones creating what I call the extended edge. And if you add the retail sector complexities increase exponentially. Think Amazon, Starbucks, Spotify and others where consumers transact via their mobile phones.

How are current hardware-only suppliers (read: Cisco) helping integrate all those devices ‘seamlessly’, a concept Cisco coined in the 90? This is now a colossal conundrum for most CIOs. On one hand, any CIO worth his or her stripes can recite in their sleep the speeds and feeds of the network, servers and storage capacity, but most have no clue as to the number of apps, version releases, usage data and productivity benefits of each. Their problem is analogous to consumers’ apps issues. As you are reading this can you honestly state how many apps there are on your mobile, their respective usage data and release versions?

So what’s the answer? More bandwidth - 5G followed by 6G, of course. The only problem is what if you only need it temporarily or periodically. Unfortunately bandwidth is not a utility (yet) but UD-WAN could potentially change that, a concept not favored by Cisco and its complicit carrier cartel!

And when we use our apps there is no service provisioning, network management, security constraints, firewalls, policies to consider, nothing in the form of the expensive and intrusive network administration. And there are no network-administration dictators controlling how we access our apps. It’s all there but the users are abstracted from those complexities. The end result is users simply download and use apps as they need them and when they need them. Apps-on-demand if you will, and it’s coming to an enterprise near you, I assure you.

Innovative SD-WAN players such as Aryaka, Breeze Networks, QOS Networks, VMWare (via its Velocloud acquisition) and a host of others are already laying the foundation for the future of enterprise networking because they know it will be applications-on-demand (user) centric and NOT box (router) centric. No more NAT, DHCP, MPLS and all that Layer 3 crap. According to Frank Cittadino, CEO of QOS Networks: “SD-WAN is currently addressing 2018 issues, which is Layer3 (routing) focused; however, the true promise of SD-WAN will be fully realized by 2021 when we focus on Layers 8/9 when we address UX [user experience] and tackle business needs with both urgency and efficiency.”

Yes, the cloud is going to be huge because we are finally ready to move up the protocol stack in the enterprise world and use software-defined services like apps-on-demand - user-defined, application-aware networking. Definitely something to look forward to! And, for argument’s sake, I purposely stayed away from including IoT (Internet of Things), which will make the cloud look and feel obese.