SD-WAN’s enterprise momentum is untethered, as more businesses look to the next-gen technology as the answer for better — and less expensive — software-defined networking (SDN). Analysts at IDC foresee enterprises’ branch network connectivity requirements continuing to drive robust SD-WAN adoption, with a predicted market growth of 70% year-over-year through 2021. As a solution, SD-WAN promises simple and powerful cloud-based management tools, nimble network policy controls, substantial cost savings through the use of commodity internet and hardware, and an ease of use and deployment that enables set-it-and-forget dexterity when establishing connected branch locations.

At the same time, SD-WAN technology is still to evolve into full maturity, and the market is still very fragmented — with every single branch networking vendor re-positioning into this space. Given the absolutely critical needs of enterprises around the performance and availability of their branch network infrastructure, business leaders should find that proceeding with prudent caution and open eyes is more appropriate than leaping headlong into adoption, tempting as it might be right now.

Cloud Provisioning Makes Deployment Easy – But There’s Complexity Hidden Underneath

When it comes to performing the initial setup and installation of network connectivity at a branch location, SD-WAN immediately distinguishes itself from traditional MPLS and CLI-managed branch router solutions by automating the provisioning and configuration of the router. Whereas the previous method of deployment called for technical staff with specialized certifications to be deployed to a location, now connecting an SD-WAN appliance to an available internet connection (and power supply) is all that’s necessary to stand up a new site if all goes well.

That said, establishing that ISP-provided WAN uplink is a vital necessity, as is ensuring that the WAN router or modem is correctly configured to enable no-hassle cloud provisioning (with the DHCP and DNS settings of the SD-WAN appliance configured automatically). SD-WAN offers the valuable flexibility of allowing enterprises to utilize the most advantageous internet provider for each connected location, as opposed to enlisting a single provider across the map even if the service is weaker for some branches. Unfortunately, this silver cloud has a dark lining: The complexity of such multi-provider networks increases the risks that WAN uplinks will not be properly configured for SD-WAN at each site. This wrinkle makes it necessary to then audit configuration settings at each site ahead of SD-WAN deployment, or to implement an alternative redundant means of connectivity to ensure reliable internet access. As versatile and advantageous as the ability to provision and manage SD-WAN over the cloud may be, it can only function if connectivity to the cloud is available and effective.

SD-WAN’s Management Advantages Depend On Maintaining Access

One of SD-WAN’s key features — and not coincidentally the reason for its name — is its use of SDN to separate the control and management plane from the data plane. Doing so makes it possible to centralize management and rely on management software that is fully independent from hardware. In this way, enterprises are also able to use the most affordable hardware available, and enjoy freedom from hardware vendor lock-in for the same reason.

However, while data center networks will often support these different planes with multiple separate networks and redundancy, if the SD-WAN overlay goes down the network can become unreachable, making troubleshooting or repair at a distance impossible. The layered flexibility of SD-WAN also exacerbates troubleshooting challenges in this scenario, making it unclear whether the source of the issue is with the ISP, SD-WAN overlay, appliance, or some misstep by a user. Unless another method of remote access is available, enterprises in these circumstances must pay the (usually very steep) expenses to have expert technicians troubleshoot the issue on-site in order to restore the network.

This hardship can be minimized if enterprises avoid single points of failure in their SD-WAN infrastructure. This means implementing multiple diverse methods of connectivity. Selecting multiple ISPs can be effective for achieving this — as long as they aren’t likely to both be brought offline by the same incident (for example, an accident that severs cables wouldn’t find the cabling for both ISPs side by side). Utilizing a cellular LTE strategy for out-of-band management is a sensible complement to traditional connectivity because of its very different risk profile. LTE’s immunity to issues that befall wired connectivity means that the branch location can remain online and available throughout outage incidents, yielding both business continuity and the ability to — securely — troubleshoot primary connection issues at a distance.

Take Action, But Be Cautious

As SD-WAN adoption increases and enterprises consider their opportunities to benefit from all that it offers, it’s important that company leaders deliberate over the risks and specific business requirements at hand. With those risks properly measured and mitigating by resilient connectivity, the transformative advantages of SD-WAN technology are ready to serve these enterprises well.

The business case for SD-WAN is very compelling. The ability to reduce reliance on expensive MPLS circuits and leverage internet connections provides significant cost savings, so all enterprises should be considering adoption now. While cloud provisioning and management is a great step in easing the overhead of managing WAN infrastructure, it does not change the fact that businesses must be able to access the remote site even (and especially) when things don’t work as expected. Making sure SD-WAN deployments include the resilience of separate management access is crucial to reduce the need for truck rolls and lessen the probability of expensive downtime.