A new startup is joining the SD-WAN space with an open architecture. The Israel-based company, flexiWAN, is seeking to solve some of the problems created by the crowded SD-WAN market today.

flexiWAN CEO and co-founder Amir Zmora first saw the need for an open SD-WAN architecture when he was working as a consultant for enterprises and service providers in the SD-WAN and Voice-over-IP (VoIP) space.

In conversations, he realized that SD-WAN had a few problems that openness could solve: IT managers have limited control of their networks and the elements and technologies that comprise them; and networking (especially SD-WAN) requires expertise in a number of domains. So no single vendor can provide in-depth expertise across all domains.

And while service providers have some opportunity in these two areas as they can prioritize and optimize different parts of an enterprise’s network, this requires a multi-vendor approach and they are forced to sell closed, vendor-specific services that don’t interconnect, Zmora said. This is not a valid long-term solution, as future enterprise requirements are unknown and will evolve and current services will be rendered useless.

So flexiWAN is breaking the vendor lock-in and “monopoly” that exists today with a “decomposed” SD-WAN architecture. “Enterprises and service providers were locked to use all networking applications from a single vendor. With flexiWAN, they will be able to pick and choose [the] best of breed,” said Zmora.

flexiWAN officially launched today, though it is still in the development phase. Its software is available now to only enterprises and service providers for private proof of concept (PoC), which it has already been conducting. It plans to launch an open source, production-ready version toward the end of this year.

During PoCs, enterprises and service providers will help identify bug fixes, security vulnerabilities and fixes, and test the software. And once open sourced, vendors will be able to contribute to the code and provide additional elements that plug into the system.

It will be delivered as software only, but Zmora said it will work with hardware vendors, such as universal CPE (uCPE) vendors, to provide both white- and gray-box options in the future.

Some services — namely uCPE — are starting to trend this way service chaining virtual network functions (VNFs) together on a piece of hardware. In fact, according to Gartner’s 2018 Magic Quadrant for WAN Edge Infrastructure, historically in the North American market, 60% of WAN edge deployments are do-it-yourself versus a managed service option.

Zmora said what flexiWAN is doing is beyond this.

“flexiWAN's open architecture allows for the integration of third-party elements into the core of the SD-WAN edge device. This makes it much more efficient as it runs in the SD-WAN edge device SW itself and not as a neighbor VNF,” he said. “No SD-WAN solution enables this today.”

Since the open architecture is not-yet production ready, the company is not announcing the third-party integrations that will be available. However, Zmora noted that it will be able to include technologies such as deep packet inspection, security, WAN optimization, session border control, VoIP, and other traffic-specific optimization logic.

The company also plans to integrate with third-party application-specific detection and optimization logic, enterprise-specific traffic and network management requirements, and open source DevOps tools.

Service providers will be able to deploy this without being tied to a black-box, Zmora said. Instead they will have full control over the system. Other providers, including managed service providers and application providers, will be able to differentiate more clearly by having closer control over how their traffic is handled.

flexiWAN claims that it is ushering in the second wave of SD-WAN.

“I think this is a revolution to the SD-WAN,” said Zmora. “It happens to be open source, but the big thing is the revolution on breaking the monopoly there — Open source really opens the door for democratizing this market.”

SD-WAN has traditionally been an area of networking where open source has not yet truly reached. “There are open-source pieces out there for VPN, FW, IDS (all the security elements), routing, and perhaps LF Networking DANOS will emerge as a viable platform on which to build uCPEs, but we haven't seen anyone try to pull a solution together yet,” said Roy Chua, founder and principal analyst at AvidThink.

Chua noted that enterprises overall might not care about an open source SD-WAN, however it could be valuable to managed, cloud, and communications service providers if it's mature and robust. So that “they are willing to take it in and try to build customized, differentiated solutions around it. It will also depend on their read on the long-term viability of the project and codebase, as well as the license that the open-source is provided under,” he said.

However, Zmora believes there could be significant market acceleration from open source SD-WAN. Not only would it excel innovation, it reduces barriers to entry and lowers development and total cost of ownership.