The changing value proposition of SD-WAN

Software-defined wide area networks (SD-WANs) have garnered a tremendous amount of interest from companies both large and small as they can significantly lower the costs and complexity of running a WAN.

As businesses migrate applications to the cloud, they are increasingly embracing the cost advantages of broadband connectivity to connect users to applications. This is being driven not only by the high cost of private WAN circuits, but because backhauling applications traffic to the data center is negatively impacting application performance, resulting in frustrated users and sub-optimal productivity.

The combination of high costs and poor performance seems like a perfect recipe for market disruption.

Cost saving has been the 'low-hanging fruit' for SD-WANs from the technology's inception. Organisations have slashed the cost of connectivity by adopting a wide range of broadband options, including consumer broadband and 4G LTE.  Also, the addition of broadband and subsequent redirection of best-effort traffic can help businesses to push out having to upgrade the MPLS circuits.

It's rare that we see companies cut the amount of bandwidth they've purchased, but they certainly could by augmenting their MPLS circuits with broadband.

I've never liked cost being the primary driver for anything, as it has limited value. Most organisations could save money on network connectivity by just renegotiating with their MPLS provider a bit harder at renewal time. Also, why go through the work of re-architecting an entire network just to save a few shekels?

For large companies with lots of trans-Pacific traffic, the cost savings can be significant, as much as 90% in some cases, but even with that I think there are more important reasons to make the shift to an SD-WAN.

It appears that customers are now thinking that way as well. In the most recent ZK Research/Tech Target WAN Survey, we asked the respondents what their top purchase drivers were for SD-WAN. Historically, cost savings was always the top response, but this year it fell to third behind the 'Need to increase WAN agility' and 'Need to shorten provisioning time for new WAN links'.

Agility and speed driving SD-WAN deployments

To me, this signals two important things. First, and most obvious, is that customers are finally looking past cost savings and thinking bigger picture: looking beyond dollars and cents, which can be difficult to do. With any new technology, it's easy to make a justification based on cost savings, but ultimately the new technology must be able to do things the old stuff couldn't.

Consider the move from TDM to VoIP. Initially, most of the deployments were done so companies could save money by consolidating two networks down to one. Eventually we figured out we could do several different things such as four-digit dialling across the globe, least-cost routing, soft phones, etc. VoIP not only saved money, but also allowed the company to do things it couldn't do before.

Similarly, an SD-WAN brings a level of agility to the network not seen before, enabling network managers to do cool new things like orchestrate network changes centrally in alignment with application requirements, shift to active-active architectures, make network segmentation easy to implement, or ove to a thin branch where all the previously resident branch infrastructure has migrated to the cloud or a regional hub.

This change is also reflected in the features that network managers are looking for from their vendors.  The top three responses in the previously-mentioned survey are 'Dynamic WAN bandwidth aggregation', 'Real-time traffic monitoring' and 'Automated network provisioning'. These all point to network managers saying, 'give me a network that is more agile and one that is easier to manage'.

There's one more important aspect to this. Based on my research, 77% of businesses surveyed are in the process of deploying an SD-WAN or have it on the roadmap to be started within the next two years. Most of these organisations are thinking about their networks more strategically.

The few businesses that are still on the fence about building an SD-WAN, should be asking themselves why? Based on my most recent survey data, the leading companies have already moved past cost savings. What's holding you back?

Zeus Kerravala is the founder and principal analyst with ZK Research.