• Underlying SD-WAN and virtual network functions address enterprise IT’s desire for more efficient operations.
• Managed service providers can differentiate themselves on operations and should invest in doing so.
I’ve heard some commentary that SD-WAN will become a feature of routers and firewalls, and the product segment will wither and die, but I don’t think SD-WAN products are going to disappear. Some vendors will consolidate features and functions. Cisco and VMware will integrate their SD-WAN acquisitions with their other networking products but that doesn’t mean enterprises will run to consolidated products. There will be some uptake of consolidated products, just like there was some uptake of router-firewall-VPN combinations, but enterprise demand for stand-alone SD-WAN products will continue for some time. There are a few reasons for this. The first is that enterprise IT likes the flexibility of bespoke products. While management is more complicated with multiple products (for those fearing for their jobs, that’s a benefit!), the upside is that enterprise IT can select a SD-WAN product that is better aligned with their requirements and allows them to replace the product in the future with a competitor product or an entirely new technology. For example, many companies were able to replace border routers with firewalls that performed the same routing functions.
The second reason is that the shift to software-defined everything is well underway and nearly all SD-WAN products have virtualized instances available for purchase in a VM format. I’d be surprised if SD-WAN vendors didn’t start shipping container formats in 2018. Where enterprises can gain advantage is consolidating network functions onto fewer appliances using virtualized formats of their favorite products. Network function virtualization is well-trod ground, following along the same path as server virtualization. Numerous vendors such as Array, Cisco, Juniper, Nuage, Versa, and VMware, are already shipping appliances that provide the underlying virtual infrastructure, package management, and service chaining—VNF connectivity—and can host third-party VNF’s. These are primarily hypervisor based systems, but I expect container formats will also be widely supported.
Finally, I think enterprise IT likes what is familiar and is loath to replace products that offer critical capabilities with another product that might be less capable overall. Rather, enterprises have demonstrated that they are more willing to take on technical debt with many bespoke products than to consolidate onto fewer that might end up being less featureful and ultimately means having all of their eggs in one vendor’s basket. Enterprises like variety and software-defined everything makes that easier.
What enterprises want—what they would pay for but will likely never get—is an environment of deep management integration across multiple vendor products which could ultimately reduce operational overhead, unlock more efficient workflows, and generate significant operational cost savings along with way. Imagine a workflow where IT defines the access policy for a use, a role, a location, a group of devices, and the system performs the configuration across all devices, either in a single vendor or multi-vendor environment. It’s a difficult task, but which vendor wants to step-up and enable their competition?
Here’s where managed service providers have a unique advantage, provided they dedicate the resources to creating a portal that integrates the management functions across vendor products, allowing enterprise IT to make changes without having to drill into each product individually. Streamlining operations is a significant value-add that enterprises themselves can’t afford individually and provides a significant incentive for enterprises that are used to building IT themselves to move to a managed service.
That’s what enterprises want.