Wednesday, May 24, 2017

Finally, a big acquisition happens in SD-WAN

One thing Silicon Valley is known for is the frantic pace of mergers and acquisitions. Cisco's recent agreement to pay $610 million for Viptela, a start-up based in San Jose, is a big sigh of relief for the many innovators and investors who have been patiently advocating a new generation of software-defined wide area network (SD-WAN) technologies for the past five years. The deal was not unexpected and one could reasonably ask, why did it take so long?

While Viptela was not an Insieme-style Cisco spin-in, the company certainly has a Cisco pedigree. Viptela was founded in 2012 by Khalid Raza, a former Distinguished Engineer at Cisco and widely regarded as a visionary in routing protocols, and Amir Khan, who had previously led the enterprise routing business and product management for MX and M series routers at Juniper. Before Juniper, Khan served as director of product management at Cisco. It is worth noting that Raza holds patents in Border Gateway Protocol (BGP) and Open Shortest Path First (OSPF). Additional founders include Atif Khan, Venu Hemige and Ramesh Prabagaran.

Interestingly, earlier this year Viptela named Praveen Akkiraju as its new CEO, replacing Amir Khan, who continued as president and board member. Akkiraju previously was CEO of VCE for the past four years, where he led the converged infrastructure provider to the No.1 market share position. Prior to VCE, he spent more than 19 years at Cisco, including lastly as SVP and GM of Cisco's Enterprise Networking group. So clearly the Viptela team will fit in well to the Cisco culture.

In May 2016, Viptela raised $75 million in its Series C, bringing total funding to nearly $110 million. The round was led by Redline Capital with participation from new investor Northgate Capital and existing investor Sequoia Capital. Proceeds from this final funding round was intended to scale sales, marketing, technical support and R&D. At the time, the company cited deployments by more than 25 Fortune 500 enterprises as well as licensing partnerships with Tier-1 carriers Verizon and Singtel.

Even with this $610 million transaction, the venture sharks may not be fully satisfied. The price should yield a 5x or better return on their investment, but it is not the billion-plus level of mergers seen in the past, back in the days when all a start-up really had to do was prove that its unique mousetrap was being deployed by a top-tier customer such as Verizon. From the seller’s perspective, some analysts have said the last funding round implied a valuation that would put Viptela in the billion-dollar unicorn club; from the buyer's perspective, Viptela appears to be in the early stages of real, recurring customer revenue generation. There are also numerous other SD-WAN start-ups looking for a buyer, and Cisco was an equity investor in VeloCloud, a direct Viptela competitor also located in Silicon Valley.

What is driving SD-WAN

The traditional enterprise WAN market brings in approximately $45 billion per year worldwide for VPN and MPLS connectivity but has not really changed for years in terms of its service portfolio, making it ripe for innovative disruption. Hence, the venture capitalists became interested in start-ups promising a soft-defined connectivity solution that can be delivered at lower-cost than MPLS and without requiring expensive infrastructure investments up front. A lot of venture funding has flowed to these start-ups for the following good reasons:

  • Businesses have an unsatisfied need WANs at a reasonable cost.
  • The emergence of cloud services means more data is going off-premises.
  • The belief that software rules.
  • The belief that the era of fixed configuration hardware appliances.
The following is a list of pure-play SD-WAN start-ups still on the market:

1.  Aryaka (Milpitas, California)

2.  Cato Networks (Tel Aviv)

3.  CloudGenix (Santa Clara, California)

4.  Cradlepoint (Boise, Idaho)

5.  Cybera (Franklin, Tennessee)

6.  Fat Pipe (Salt Lake City)

7.  Glue Networks (Sacramento, California)

8.  Mushroom Networks (San Diego)

9.  Silver Peak (Santa Clara)

10.  Talari (San Jose)

11.  Velocloud (Mountain View, California)

12.  Versa Networks (Santa Clara)

13.  Webscale Networks (Mountain View, California

Established network hardware vendors also offer SD-WAN solutions. Even if they did not originally use the term SD-WAN, many are doing so today. Prior to the Viptela announcement, Cisco already had its iWAN network overlay solution for enterprises and service providers, and its Meraki group also offers a cloud-based WLAN management that could also be considered a software-defined network management service.

Then there is Riverbed, which is normally thought of as the original WAN optimisation company. This strategic position fits in well with SD-WAN. In 2016, Riverbed acquired Ocedo Networks, a developer of SD-WAN technologies based in Karlsruhe, Germany.  Riverbed had been using Ocedo to power its Project Tiger initiative of application-centric SD-WAN solutions designed to eliminate the need for traditional branch routers. Last month Riverbed agreed to acquire Xirrus, a leading supplier of enterprise and service provider WiFi solutions. The combination of these two newly acquired companies, along with its link optimisation technology, should make Riverbed a potent player in this space.

Many can play this game

Network aggregation technologies have been around for a long time. While independent service providers have become quite successful at aggregating access lines from multiple carriers into a software-driven WAN as a service. Global Capacity is a good example. Since 2000, this carrier based near Boston has been building virtual networks for enterprises. Its One Marketplace hub aggregates the access and pricing information for broadband telecommunications services, including MPLS VPN, Ethernet over copper and DSL. In 2014, Global Capacity acquired MegaPath’s network services business unit which include the Covad DSL assets.

A similar company is Virtela, which NTT Communications acquired in 2014. The Virtela proposition was also simple: enable the enterprise customer to build a virtual overlay network from multi-carrier MPLS, Ethernet, DSL, 3G/4G/LTE and other IP links, while benefiting from a single SLA and management portal. Virtela operates global operations and delivery centres in the U.S., India and the Philippines.

In addition, at least two major service providers (AT&T and Orange) are building their own SD-WAN solution rather than buying a solution from one of the start-ups or an established vendor. So the question of which technology provider will dominate the SD-WAN segment is still in the open.

Concluding thoughts

Now that Viptela is off the field of potential acquisition targets, will it be easier or harder for the remaining start-ups to find a suitable dance partner? In the past, whenever Cisco bought a start-up in a particular category one could expect Juniper, Ericsson or Alcatel-Lucent to do the same. But this dynamic has not been seen for several years.

One could still imagine Juniper buying up Versa Networks or HPE making a bid for Velocloud or Webscale Networks; and Ericsson might find a new focus and liking to this segment. The most likely buyers, in OND's opinion, would be the giant cloud companies, especially AWS, Microsoft and Google. These companies clearly benefit when new railroads are built to bring traffic into their domains. As cloud providers, they are already delivering the infrastructure for running the controllers used by these SD-WAN start-ups. Managing their own SD-WAN could also provide a cloud company with a competitive advantage and a new billing opportunity. Even for smaller companies looking to provide secure Office suite software to multiple branch offices, why not buy an integrated WAN connectivity directly from Microsoft? So the question is what are they waiting for?