WAN connections are complicated, expensive to maintain and difficult to change, inhibiting rapid innovation and growth. Last year I proclaimed the SD-WAN market would face consolidation as software-defined WANs (SD-WANs) gained momentum. Since then many have confirmed this view, adding that market consolidation would accelerate once networking giant Cisco Systems made an acquisition. The wait is over as Cisco recently acquired Viptela, a leading SD-WAN vendor. Cisco already had one of the most deployed software-based WAN products, iWAN, but few customers who I have spoken to are enamored with it. In fact, an entire ecosystem has grown up around making iWAN easier to use. We have written about Glue Networks, for instance, who has built orchestration around iWAN to help reduce complexity and enable IT to be more agile. SD-WAN, unlike iWAN, is significantly easier to deploy and manage.
Cisco buying Viptela is the first real shot in the battle, a recognition that SD-WAN is not a fad, is going to be a real business and can no longer be ignored. SD-WAN has momentum and customers love it. Most importantly, Viptela is a pure software play, so Cisco can offer this to its existing customer base on top of its existing hardware products without any issues. I see this as a positive for Cisco, finally admitting that the current WAN structure was untenable. In a world of agile businesses, a flexible software layer on the top can alleviate WAN complexity and let the businesses move faster. John Vincenzo, SVP of Silver Peak, stated in an email to me,
“The Cisco acquisition of Viptela serves as a clear indication that pure-play SD-WAN vendors that can’t address the broader performance and security requirements of the WAN edge will be challenged to build stand-alone businesses, and are likely to be consolidated.”
I am primarily in agreement of his assessment, but put another way, if you are only a software layer, the barrier to purchase and integration is a lot lower. Unfortunately, so is stickiness and switching costs for your customers, making it easier for a business to jump ship if you are not innovating. Part of the appeal of a pure SD-WAN play for customer is the very thing that creates challenges for vendors in growing and sustaining the business: lack of a consistent ongoing revenue stream.
SD-WAN technology is reaching market velocity, and deployments are coming along quickly. But most of the 30-40 vendors deploying SD-WAN are simply touting the ability to create a hybrid WAN, basically a WAN environment that uses multiple connections (like MPLS, internet, 4G LTE or others). This is an interesting feature, but long-term differentiation will be difficult. It is hard to build an empire on a feature, which is why most of the dozens of businesses will fail. The evolution of SD-WAN will be Company (yesterday) → Product (today) → Feature (near future) → Table Stakes (long-term future). In a few years SD-WAN will be an assumption, not a differentiation, but until that happens, expect a lot of jockeying in the market.
After a year of watching the market, talking to customers and studying the technology, I believe we’ll see the following three major trends:
- Consolidation: Nuage went early, consumed by Nokia, who has continued to operate it with much autonomy; Viptela was second in line, moving under the Cisco umbrella. The better companies in the market are probably now all in play: CloudGenix, VeloCloud, Versa Networks and Silver Peak are probably top contenders here. They have good technology and more importantly a market footprint. I expect equipment vendors to be prime suitors, but don’t discount a carrier, MSP or large cloud provider stepping up somewhere.
- Cloud Connectivity: Early SD-WAN offerings were pitched for hybrid WAN connections, reducing MPLS costs, and branch office automation. Today the big pitch for SD-WAN is as a cloud connectivity layer for datacenters. Many are now offering specific cloud optimizations and even “cloud traffic reports.” Clearly, they see branch office cloud connectivity being a bigger opportunity than just connecting to HQ. Routing web and cloud app traffic directly out of the branch and not hair-pinning it through the HQ datacenter and back is a big cost saver and simplifies traffic management greatly—as long as you have a good hold on the security.
- Carrier Deals: While enterprises are a bigger market (with higher ASPs), they move slowly; the real action in the past year has been carrier deals. SD-WAN vendors are locking down their footprint with carriers and that could be the Trojan horse for getting into enterprises in the long run. If your carrier is already using product X to connect you, it creates a great bias toward purchase when your enterprise decides to start down the SD-WAN path.
All of these trends point to good future for SD-WAN. But that is not an infinite path. Eventually all networks will have software definition, because hard coding the connections and rules is time consuming and inflexible. By this time, SD-WAN will simply be called WAN, as the software-definition of the whole network will mostly be expected. Much like seatbelts—which were once an upsell or an advertised feature—are now standard equipment expected in every vehicle. Eventually, SD-WAN will be just as mundane. But until then, expect a lot of movement in the market.