SD-WAN is no longer a new technology. Over the past several years, it has evolved from a way of simplifying connectivity between sites into a broader platform for connecting users, applications, cloud services and security environments. But maturity does not automatically mean simplicity.
For organisations reviewing SD-WAN in 2026, many of the same questions remain: what problems should it actually solve? Where do deployments go wrong? How should organisations think about underlay connectivity? And what separates a deployment that delivers long-term value from one that simply recreates a traditional WAN with a new management interface?
We asked one of our SD-WAN experts for their perspective on how the technology has changed, the biggest challenges organisations are trying to solve, the misconceptions that still exist, the lessons learned from deployment and what IT leaders should consider when reviewing their WAN strategy in 2026.
Let’s get into it.
Q. How would you explain SD-WAN to someone who understands networking but hasn’t worked with the technology before?
A. SD-WAN is a technology that allows connecting offices, data centres, cloud services and remote locations through a centrally managed network independent of the underlying transport technology (MPLS, mobile, satellite, DIA, broadband, etc).
Q. Looking back over the past five years, what have been the biggest changes in SD-WAN?
A. The biggest change in SD-WAN over the past five years has been the increase in native support and integration with external 3rd party services such as SSE providers (Zscaler, Palo Alto, Cloudflare, Cisco, Fortinet), SaaS providers (Microsoft 365, Webex, Google Workspace) and cloud service providers (Azure, AWS, GCP). Early implementations of SD-WAN focussed primarily on connecting customer’s on-premises sites and permitting internet access in a secure and efficient way. However, connecting directly to 3rd party services was often a supported, but manual, procedure that required the creation of tunnels, complex routing policies and custom monitoring services. This native integration provides automated configurations and simple, controller managed, deployment of 3rd party services across all customer sites.
Q. What are the biggest challenges organisations are trying to solve with SD-WAN in 2026?
A. Two challenges continue to stand out: reducing WAN transport costs and eliminating unnecessary traffic backhaul.
Reducing WAN transport OPEX
Because SD-WAN is transport agnostic and provides secure connectivity between customer locations by default, there is less of a dependency for private WAN transports (MPLS, VPLS) which often have high costs when compared to public transports such as the internet. Generally, when migrating to SD-WAN, customers are also looking to terminate contracts with their existing private WAN providers and switch to significantly cheaper contracts with ISPs. Additionally, due to SD-WAN’s ability to detect and mitigate poor performing circuits in real-time, the availability and performance SLAs of the WAN transports do not necessarily need to be as strict as in traditional networks and therefore cheaper WAN circuits can be purchased. Again, reducing WAN transport OPEX.
Reducing unnecessary backhaul
SD-WAN allows organisations to securely use local internet breakout (also referred to as direct internet access or DIA) mitigating the requirement to backhaul internet-bound traffic to central locations, whether data centre or colocation. Traditionally, traffic would be backhauled either because only private WAN connectivity exists at a remote site or for security inspection on a centralised firewall. The introduction of internet as a WAN transport at remote sites and the NGFW capabilities built into SD-WAN appliances allows traffic to traffic to breakout locally to the internet. This normally reduces latency for users connecting to internet applications/services and reduces the throughput demands on centralised equipment often leading to reduced power consumption, licensing costs and hardware expense.
Q. What misconceptions do you hear most often about SD-WAN?
A. Any WAN underlay transport can be used regardless of the business requirements.
Whilst SD-WAN can overcome and mitigate WAN circuits that are temporarily underperforming (outages, latency spikes, packet loss, jitter variance) it’s performance potential will always be bound by the maximum capabilities of the highest performing underlay transport. For example, if a remote site has two internet circuits; circuit A with a RTT to the organisation’s data centre of 50ms and circuit B with a RTT of 65ms the best latency performance that SD-WAN can achieve will always be 50ms. Some business applications and services such as call centres or trading platforms may have strict performance requirements that include latency <45ms. In this example, those requirements will be violated with “standard” internet circuits. The type and quality of underlying transports must always be mapped to the business requirements of each site.
Q. Where do organisations go wrong when planning or deploying SD-WAN?
A. One common area where organisations fail when planning for and deploying SD-WAN is the standardisation of the existing LAN network at each remote site. One of the advantages of having all SD-WAN appliances managed by a centralised controller is the ability to template and standardise the WAN. However, the configuration of SD-WAN routers is intrinsically linked to the architecture of the site’s LAN where, for example, a template has a fixed number of interfaces. When a site follows a router-on-a-stick architecture, the number of sub-interfaces configured on the SD-WAN router is equal to the number of layer 3 VLANs at the site.
Additionally, the names of the sub-interfaces will be dependent on the site’s VLAN numbering schema. If each remote site has a different VLAN architecture and/or different numbering schema then each site will require its own SD-WAN template. Thus, defeating the point of using templates for standardisation. This is just one example of where a non-standardised LAN can impact the ability to standardise the SD-WAN. Careful planning and remediation of significant LAN variances in advance of an SD-WAN deployment is essential.
Q. What separates a successful SD-WAN deployment from one that struggles?
A. Successful SD-WAN deployments are usually ones that take full advantage of native features available from the vendor and don’t simply lift and shift configuration, such as routing and security policies, from a traditional network onto an SD-WAN one. There can be a temptation during the design and migration phases of a deployment to simply recreate existing configuration in SD-WAN, however, whilst this can be an easy to understand and quick migration method, it results in an SD-WAN network that operates in a very similar way to a traditional network but with a controller and a modern GUI. Examples include:
- Policies that match traffic based on IP addresses and port numbers as opposed to built-in application recognition.
- Manual and complex BGP route policies as opposed to SD-WAN rules/ application aware routing and automatic topologies.
- Manual IP SLA for tracking WAN status rather than native SD-WAN tunnel performance monitoring and SLAs.
Q. What advice would you give to an organisation reviewing its WAN strategy in 2026?
A. It is important, and often extremely beneficial, to understand in detail, the users, applications, and services operating within your organisation. From an WAN design perspective, examples include:
- Number and types of users and clients at each site
- Security classification of each type of user/client (e.g. trusted, untrusted)
- Average & peak throughput for each site
- Applications consumed by each user/client type
- Classification of each application’s business importance and any SLA requirements
- Hosting location of each application (on-premises, cloud, internet)
- Security classification of each application (e.g. trusted, untrusted)
Having a detailed understanding of the above information, allows an organisation to make fully informed decisions about WAN strategy that are appropriate for their existing, and future, network. Additionally, the above information can unlock the full benefits and potential of an SD-WAN deployment, improving security controls, reducing operational complexity, and improving performance and reliability of the network.
SD-WAN in 2026: Start with the requirements, not the technology
SD-WAN has matured considerably, but many of the lessons behind a successful deployment remain surprisingly consistent. The technology can simplify operations, provide greater transport flexibility and improve how organisations connect to cloud, SaaS and security services.
But SD-WAN does not remove the need for good network design and SD-WAN management.
The underlay still matters. LAN standardisation still matters. Application requirements still matter. Simply migrating an existing configuration into a new platform rarely delivers the full value of the technology. Ongoing optimisation and management are just as important, ensuring the environment continues to perform effectively and delivers as much value from the investment as possible.
For organisations reviewing SD-WAN in 2026, the most important question is therefore not simply: “Should we deploy SD-WAN?” It is: “What does our business need the WAN to do, and how can SD-WAN help us deliver that more effectively?“
Understanding that first gives organisations a much stronger foundation for reducing unnecessary cost, improving performance, simplifying operations and building a WAN strategy that supports both today’s requirements and what comes next. Contact us today to schedule your technical SD-WAN workshop.