The announcement of VMware’s intention to acquire VeloCloud signals the broadening of the NSX Everywhere story. SD-WAN is a solution that offers agility, security, orchestration, and other business outcomes for remote and branch offices. It should not be considered just an MPLS replacement for the WAN with savings on bandwidth costs.
At a core level, both NSX and VeloCloud’s products are based on an overlay network, which offers the flexibility to treat a logical network separately from the physical network, and this core concept has been popularized for many years via MPLS. Ironically, it’s the perceived lack of flexibility and costs of MPLS that have become the initial drivers for the popularization of SD-WAN, which promised to modernize the branch networks and WAN.
VMware’s NSX Everywhere plan is similar to Cisco’s ACI Anywhere plan since it enables the core data center networks to reach out into other locations such as a public cloud.
There are interesting possibilities ahead. Cisco is the leader in branch routers, via devices such as the popular Integrated Services Routers (ISRs). Those devices will evolve to offer more capabilities via virtual network functions, as the latest routers are platforms in which to run VNF. So Cisco is becoming more software-centric.
VMware is traditionally a software-only company that does not sell devices. Does it aspire to start to offer branch devices because VeloCloud Edge is such a hardware device, or place the functionality into virtual network functions (VNFs) that can be deployed within vSphere at the branch office? The latter makes more sense. There may be a transition period in which VeloCloud Edge device is offered, though, since ultimately the VeloCloud Edge is, underneath the cover, an x86 network device that can run as a VNF in a VMware-based edge hypervisor, which is a general networking VNF platform at the edge. This is part of VMware's NFVi platform vision.
Ultimately, the vision of VMware is to create a software-defined network that spans the enterprise, and VeloCloud brings them closer to the vision. VMware has the technical assets to create an SD-WAN solution on its own, but this acquisition is a quicker way with gained access to VeloCloud’s technology and customers (enterprises and service providers alike).
There also are two parts to the equation: the edge and the cloud. First, there is an opportunity for VMware to provide SD-WAN VNFs at the edge, running on commodity servers. Secondly, there is an opportunity for service providers and telcos to offer the SD-WAN cloud service on their own telco cloud. VeloCloud used a public cloud provider for that, but telcos are probably eager to run that in their telco clouds as a way to differentiate via better SLAs or with value added services.
What about WAN vendors?
Another part of the equation is the other SD-WAN vendors, such as Riverbed and Silver Peak who expanded their heritage of WAN optimization to become SD-WAN vendors. Does the acquisition of Viptela by Cisco and VeloCloud by VMware shine a spotlight on Riverbed and Silver Peak as pure WAN vendors with a singular WAN & ROBO focus? Or do the larger set of resources and assets of Cisco or VMware steal attention? A hidden story may be that it may renew focus on the edge functions and services delivered as a set of software services in VMs.
Riverbed has a SteelFusion product, which is a hypervisor-based system that converges many branch services into one box. This is the vision that VMware or Cisco may be pursuing after all on the long term, and we may see a convergence of the visions of many major players into a similar architecture.
SD-WAN Is an Edge Access Story
This also puts more attention on Citrix NetScaler SD-WAN, since Citrix has a strong position on remote access to apps (ADC), files, networks, and desktop workspaces (XenApp, XenDesktop in data centers and cloud) and ultimately SD-WAN is an remote resource access solution, not a pure standalone network solution, so a firm with a wide portfolio can offer a comprehensive solution that a customer needs. SD-WAN is a glue that delivers services to the edge, and Citrix is also a company that prides itself on being an enterprise glue, so its solution can gain more attention. This is related to a concept of an enterprise service mesh that I am exploring, which Citrix is delivering on.
Many Parties in the Fray
Let’s not forget the slew of other companies, big and small alike, that offer some type of SD-WAN product. How they will work together, independently or in partnership with other established players, is a big question.
There are many of them, including Aerohive, Aryaka, Barracuda, Big Leaf, Cato, CloudGenix, CradlePoint, Cybera, Ecessa, Elfiq, FatPipe, Fortinet, Glue Networks, Huawei, InfoVista, Juniper, MultApplied, Mushroom, NTT / CloudWAN, Nuage, Riverbed, Sonus, Talari, TeloIP, Versa, ZeroOutage, and ZScaler (forged many CASB partnerships with SD-WAN vendors). Some are direct SD-WAN vendors, while others extend their products to offer SD-WAN or partner with other firms. There are also many telcos and managed service providers too numerous to list, but the major ones like ATT, Sprint, and Verizon in the US come to mind, and many already have SD-WAN offerings.
I suspect there will be more partnerships and other announcements to come.