SD-WAN Poised to Steal Thunder from MPLS in 2018

November 15th, 2017 by · 2 Comments

This Industry Viewpoint was contributed by David W Wang

At the end of 2017, we already see quite some momentum gathered around SD-WAN new service launches in the coming 2018 from tier one & two communication service providers, both in the US and global markets.

Over the next five years to 2021, it’s predicted the global MPLS revenue grows only at CAGR 4.1%, while SD-WAN will enjoy a robust 69.6% CAGR.

If we say 2017 still sees SD-WAN launches on trial and warm-up basis from numerous service providers, then 2018 will mark the real game kick-off for SD-WAN to take on MPLS in the competitive WAN marketplace. In many cases for the enterprise clients today, SD WAN has disruptive advantages over MPLS in terms of cost, provisioning, agility, speed, direct connections to cloud, service redundancy, and management transparency.

Some service providers, such as cable operators Comcast and Charter, will debut their SD-WAN with no legacy MPLS service in the background; while major telcos like AT&T, Verizon and CenturyLink would have to consider the “cannibalization effect” to their existing MPLS revenues as they roll out their own SD-WAN solutions as well as facing competitions from the cable operators and other SD-WAN players.

So these are the two critical questions concerned the major telcos need to address in 2018: first, to what extent SD-WAN may take market shares away from MPLS? Second, how to best maneuver the headwinds of SD-WAN competitions and keep the WAN business healthy?

Based on various forecast sources to date (from such as IDC, Grand View Research, Research and Markets, etc.), in 2018 the global MPLS revenue may reach $42.53 Billion, and SD-WAN revenue would hit $2 Billion. The assumption is most of the $2 Billion for SD-WAN would be from the migration of existing MPLS services although at the first sight, this impact seems to be not that significant.

But we need to keep one key caveat in mind: a SD-WAN site service can go 5 to 8 times cheaper than the MPLS counterpart, depending on the particular site circuit status, access configuration, and class of services (CoS). This means when a client migrates their WAN service from MPLS to SD-WAN, the MPLS service provider will either lose the account to other vendors, or erode the current revenue to a much lower level despite of the rescued client’s loyalty.

Therefore $2 Billion revenue gain for SD-WAN in 2018 might indicate potentially $10-16 Billion MPLS revenue erosion for major telcos, which becomes quite a serious wake-up alarm.

What are the best strategies to handle such looming MPLS revenue cannibalization? There seems to be not a short and simple answer to this challenging situation. The telcos have to do more homework and dive into details on the SD-WAN deployment and competitions in their service territories, and also the impact on different customer use cases, before they can decide on what to do next.

From the angle of service deployment, SD-WAN can go in three scenarios. The first scenario would be using pure public Internet (most likely with redundancy from two different ISPs), which means the client would totally unplug the existing MPLS service and switch to high speed SD-WAN with intelligent routing and security package over the public Internet backbone.

In the second scenario, the client may subscribe to SD-WAN over the public Internet but still keep their MPLS service. Via such a hybrid solution hopefully with dynamic load balancing and fail-over, the client may route non-mission critical traffic like software usages, email, and storage file transfers over SD-WAN, while still handle mission critical applications like voice and real-time video over MPLS.

As the result, the client may downsize their MPLS bandwidth for cost saving purpose. For instance, if the site used to take 10 Mbps MPLS service, now the client may reduce it to 5 Mbps for handling the mission critical traffic only.

The third scenario is not very common yet but surely worth attention. Some service providers may offer to put SD-WAN traffic over their own private layer 2 backbone, or a hybrid setup with a public Internet access. In this way, their client while leaving a MPLS service provider may still enjoy the CoS and security guaranteed for WAN communications, especially for mission critical traffic.

While all the three scenarios above could hurt the existing MPLS services, still major telcos should strive for the second scenario, so as to at least prevent the immediate accounts churn and keep the chance alive to make up for the MPLS revenue erosion through developing new value-added services.

From the customer use case perspective, we also see three major models, considering the end user’s traffic and application types as well as geographic service coverage.

The first model is for metro and regional level, due to the enhanced Internet quality from dense fiber optic network deployed in recent years, a qualified SD-WAN solution, once tested well, should be capable to replace existing MPLS services, regardless of the traffic and application type being mission critical or non-critical.

If the WAN service scale is on the national level, the reliability of fiber based and meshed public Internet backbone has also been upgraded significantly.  On the safe side, however, adequate tests on the to be frequently used routes should be done first before deciding to adopt pure SD-WAN over the public Internet or use a SD-WAN and MPLS hybrid.

On the international level, the common rule of thumb is the public Internet backbone across the north America, Western Europe, and most of Asia Pacific region should be reliable and scalable enough for enterprise applications. But overall, MPLS still has lower level of packet loss, latency and jitter than SD-WAN for international long-haul transport.

A multinational enterprise client therefore may want a hybrid MPLS and SD-WAN to handle mission-critical, on premise, real-time traffic, and those cloud based, non-real-time applications respectively.

It then becomes obvious that those multinational large enterprises are the targets for major telcos to safeguard the existing MPLS revenues. This is not saying the telcos should delay rolling out their own SD-WAN services. The SD-WAN trend is unstoppable and one can only go with the flow or lose to competitions.

To better serve the multinational enterprise segment, the telcos may set up goals of using SD-WAN to augment the existing MPLS services and reinforce the customer loyalty. For example, the hybrid MPLS and SD-WAN service can be applied to more branch and remote sites (including IoT applications), and cloud applications for the enterprise clients.

David W Wang is a telecom/IT business development principal and senior consultant with ITCom Global, LLC based in Washington DC metro and author of the new book “Cash in on Cloud Computing. David can be reached at

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Industry Viewpoint · SDN

Join the Discussion!

2 Comments So Far

  • Rob Powell says:

    I have repaired the formatting on this article. I blame hotel WiFi… haha. Sorry about the inconvenience!

  • Dave says:


    Solid analysis and a particularly good point on explaining how MPLS displacement by SD-WAN may mean far more than a 1:1 replacement.

    You accurately identified the main options for service providers and enterprises and the need for MPLS in global deployments. Performance between Internet regions is most definitely a big issue. Service providers and enterprises do have another, less discussed alternative — Internet transit.

    For years, small ISPs would purchase access to Internet backbones using Internet transit services. Transit services offer more predictable access than public peering. Using an SD-WAN in the core to select between transit services across multiple backbones provides latency/loss better than the Internet at a ~1/10 of MPLS costs.

    Not only have ISPs used this method, but so have CDN players, like Akamai. Cato Network took that approach to build the backbone underlying its’ SD-WAN service. It’s a proven approach that service providers looking for an alternate MPLS-offer or enterprises looking to replace MPLS might consider.

Leave a Comment

You may Log In to post a comment, or fill in the form to post anonymously.

  • Ramblings’ Jobs

    Post a Job - Just $99/30days
  • Event Calendar