VMware’s VeloCloud acquisition: an argument for SD-WAN services?

The recent news around VMware’s acquisition of SD-WAN provider VeloCloud is puzzling from a lot of angles but particularly in what it says about SD-WAN services.

Let’s make a deal

VeloCloud is a leader, and some would say the leader, in the SD-WAN market. The company has been in the space since its founding in 2012 and has raised $84 million in private funding, according to CrunchBase. It claims around 1,000 enterprise customers (1,000).

The VeloCloud acquisition will help VMware compete with Cisco, who acquired SD-WAN provider Viptela for $610 million in May. VeloCloud isn’t VMware’s first virtual networking acquisition. Back in 2012, the company acquired Nicira, which became the basis for its NSX network virtualization offering. Integrating the two technologies creates an interesting end-to-end solution. VeloCloud’s approach of coupling appliances with aspects of a cloud service, will play well with VMware’s premise-oriented strategy.

But it’s VeloCloud service provider focus that’s particularly significant. The company has forged partnerships with 50 service providers to date. VeloCloud also delivers its SD-WAN appliances as a virtual network function (VNF) on universal CPE, a common infrastructure for service providers.

The deal “signals a strong commitment to build deeper partnerships with communication service providers (CSPs),” writes Gabriele Di Piazza, the vice president of products and solutions at VMware in a recent blog.

VeloCloud delivers its SD-WAN appliances as a virtual network function (VNF) on universal CPE and has forged service provider partnerships (50 at last count). The company also provides strategically placed “gateways” allowing for shared exit to private cloud infrastructure and traffic exchanges for VeloCloud suppliers to peer with another. All total, the VeloCloud approach provides the components enabling even regional service providers to deliver a global SD-WAN services.

Service provider SD-WAN, anyone?

And it’s the relationship between SD-WAN and carriers that has me scratching my head. On the one hand, SD-WAN appliance vendors look to service providers to fill key pieces in their SD-WAN solutions — the private networks to overcome Internet routing problems; the design, implementation and management expertise to roll out the SD-WAN; and the service level agreements (SLAs) coveted by enterprises.

“The acquisition of SD-WAN vendor VeloCloud, is an indication that WAN edge technology is rapidly commoditizing and is essentially a feature in a network service provider offering. The true battle is for the future of network services for regional and global enterprises,” Shlomo Kramer, co-founder and CEO of Cato Networks, wrote to me in an e-mail.

On the other hand, buying an SD-WAN from a carrier or other traditional MPLS providers is a strategic mistake. When carrier sell SD-WAN services they typically require customers to have at least one carrier circuit at most locations. You give up your leverage to improve the service experience. And after spending more than a decade working carriers, trust me, that’s the last thing you want to do.

There’s a reason traditional MPLS providers are among the most hated of IT suppliers. Companies need to wait weeks even months to receive an MPLS circuit. Making just simple change to a router requires opening a ticket with the service provider that can take 24 hours to fulfill. Didn’t like how the provider handled your last ticket? Too bad. The pain and cost of switching providers was so great that most midsize enterprises ultimately had to “swallow it.”

And you expect these providers to deliver a nimble service like SD-WAN? The opposite is more likely.

By owning the SD-WAN equipment you can easily switch service providers. It’s an enormous strength that we’ve rarely had before with enterprise network services. The other benefits carriers bring can be found whether through the new breed of service providers, such as Aryaka or Cato, that can use any transport, good local Internet access, and independent WAN designers for the deployments. Even the ongoing NOC/SOC monitoring is possible with third-parties.

It’s more an Internet approach, requiring the integration of a few providers for a total solution, but the net result is you retain control over your providers. When you buy an SD-WAN service you give up on that control and become susceptible to the delays and overhead SD-WAN tried to eliminate.

Why would you want to do that?