SD-WAN explained: is it right for your business?

By Paul Botwright, Director, PrimeTel · Published 10 November 2026

SD-WAN is sold hardest on the one benefit that is least reliable. The technology is sound, the savings are not automatic, and whether it suits you comes down to how many sites you run and where your applications actually live.

SD-WAN, short for software-defined wide area network, is a way of managing the connections between your sites, your users and your cloud platforms from one central place. Rather than each office having a fixed line that carries everything the same way, SD-WAN sits across several connections and decides, traffic by traffic, which one to use. The timing matters this year: the old copper phone network is switched off on the 31st of January 2027, and a lot of businesses replacing lines for that reason are making decisions about the rest of the network at the same time.

Most of what is written about SD-WAN is written by people selling it. This is the version I give clients: what it actually does, how it compares with MPLS, which benefits hold up under scrutiny, when it is worth doing, and when it is not.

In short

SD-WAN manages traffic across several internet connections from one central platform, so you can prioritise voice over file downloads, fail over automatically when a line drops, add a site in days rather than months, and send cloud traffic straight out rather than hauling it back through a data centre. It usually suits businesses with several sites, heavy cloud use or quality problems on voice and video. It usually does not suit a single-site business, and it cannot improve a connection that is simply poor. The cost saving everyone leads with only materialises if you genuinely cease the old circuits.

1. What SD-WAN actually does

Think of the old arrangement. Each site has a line. Everything that site does goes down that line, in the order it happens to arrive. A large file upload and a phone call compete, and the phone call loses.

SD-WAN puts a piece of software in front of that. Each site can have two or three connections, usually a mix of business fibre, ordinary broadband and sometimes a mobile connection as a backup. The software knows which applications matter, watches how each connection is performing second by second, and sends each type of traffic down whichever one is behaving. Voice goes on the clean path. The backup download goes on the other one. If a line degrades or drops, traffic moves across without anybody noticing.

The second half is the management. Instead of logging into equipment site by site, you get one place to see every location, set rules once and push them everywhere. In practice that means:

Voice and video get priority without anybody configuring each site by hand
A failed line moves traffic rather than stopping work
A new site can be live on a standard broadband connection in days
You can see which site has a problem before the site rings to tell you

2. SD-WAN vs MPLS: what actually changes

MPLS is a private network. Your traffic never touches the public internet, the provider guarantees how it performs, and you pay for that. The drawbacks are the ones everyone knows: a new site takes months, bandwidth is expensive, and traffic for a cloud application often goes back to a central data centre before coming out to the internet, which adds delay for no benefit.

SD-WAN takes the opposite approach. It uses ordinary internet connections, which are cheap and quick to install, and manages their unpredictability in software. Cloud traffic can break out locally instead of taking the scenic route.

The honest comparison is that you are trading a contractual guarantee for flexibility and cost. MPLS promises performance and charges for it. SD-WAN manages performance and relies on you having enough connectivity for the software to work with.

It is not either or. Plenty of businesses keep MPLS where it matters, often a head office or a data centre link, and run SD-WAN everywhere else. If you are comparing the two, ask for:

The full cost of both options over three years, not a circuit-for-circuit price
What happens to the MPLS contract: when it ends, what notice it needs, what leaving early costs
Which applications genuinely need a guaranteed path, and which only sound like they do
Who carries the performance commitment once the guarantee is gone

3. The benefits that hold up

Five benefits stand up when you look at them closely. The one nobody oversells, and everybody values afterwards, is control: one policy change applied across forty sites, rather than forty changes.

The resilience is real too. Two ordinary connections with automatic failover are usually cheaper and more reliable than one expensive line, because they fail in different ways. A digger through a duct takes out one path, not both. What holds up:

One place to see and change every site
Automatic failover between two ordinary connections
Voice and video prioritised when the office gets busy, which is a better fix than buying more bandwidth
New sites live in days rather than months
Cloud traffic sent out locally, instead of through a central site first

4. The benefit that often does not

Cost. Almost every SD-WAN proposal leads with the saving against MPLS, and the saving is real in the right circumstances. But it is a saving on circuits, and it only arrives if you actually cease those circuits. Three things get in the way.

You are usually mid-term on the MPLS contract, so the old cost runs alongside the new one until it ends.

SD-WAN brings its own recurring costs. There are licences, usually per site and per year, often with hardware that is either bought up front or rented monthly, and a management fee if somebody else is running it for you.

And teams frequently keep one MPLS link longer than planned, for one application nobody wants to move.

None of that makes SD-WAN a bad idea. It means the business case has to be built on the whole picture over three years, with the overlap period in it, rather than on a comparison of line rental. The places proposals usually go quiet:

The overlap while the old contract runs down
Licence costs per site, and what happens when you add one
Whether the hardware is yours or rented, and what it is worth at the end
The cost of the managed service, separated from the technology

5. When SD-WAN is right for your business

It tends to be worth it when more than one of these is true:

You run several sites. Below a handful, the maths rarely works
Your applications have moved to the cloud, whether that is Microsoft 365, a hosted phone system or a cloud contact centre
You have quality problems on voice or video that more bandwidth has not fixed
You open, close or acquire locations, and waiting months for a line costs you
Your existing contract is ending

That last one is the practical trigger for most of these decisions, and the reason to start looking a year out rather than three months out.

6. When it probably is not

The one worth being blunt about is connectivity. SD-WAN chooses between the connections you have. If every line into a site is poor, routing between poor lines does not produce a good one. Fix the access first. It is also unlikely to be worth it when:

You have one site, or two. A decent firewall and a second connection give you most of the resilience for a fraction of the cost
You are early in an MPLS term with painful exit costs. Plan it now and do it at the contract end
Nobody will own it. A co-managed service where neither side is clear who changes policy ends up with nobody changing policy

7. What to ask before you sign

Whoever you end up talking to, these are the questions that separate the proposals:

Is this fully managed or co-managed, and exactly which changes can we make ourselves?
What is the licence term, and does it match the circuit term? They often do not, and the mismatch is where people get tied in.
Is the hardware purchased or rented, and can we keep it if we change connectivity provider?
What does the reporting actually show us, and can we see it before we sign?
What is the response at two in the morning when a site is down, and what is the commitment in writing?
What are the exit terms, and what do we have to give back?
Who is responsible for performance when the underlying connection is not yours?

How PrimeTel helps you choose an SD-WAN provider

We do not sell SD-WAN, and we are not a reseller for any of the vendors. We run a structured comparison on your behalf: working out whether SD-WAN is the right answer at all, defining what you actually need, putting the same brief to the suppliers worth considering, and comparing what comes back on a like-for-like basis, including the costs that tend to sit outside the headline price.

If it turns out you do not need it, that is a legitimate outcome and we will tell you.

You can learn more about our approach to Data Network Management and Wide-Area Network (SD-WAN), including how we compare providers.

Frequently asked questions

What is SD-WAN?

SD-WAN, or software-defined wide area network, is software that manages the connections between your sites, users and cloud platforms from one central place. It decides which connection each type of traffic should use, so important traffic such as voice gets the best available path.

What is the difference between SD-WAN and MPLS?

MPLS is a private network with a contractual performance guarantee, priced accordingly and slow to add sites to. SD-WAN uses ordinary internet connections and manages their variability in software, which is cheaper and faster to deploy but moves the performance responsibility from a guarantee to a system.

Does SD-WAN save money?

It can, but only if you genuinely cease the circuits you are replacing. Licences, hardware and management fees are new recurring costs, and the old contract usually runs alongside the new one for a period, so the business case needs to cover three years with the overlap included.

How many sites do you need before SD-WAN is worth it?

There is no fixed number, but the benefits scale with locations, and below a handful of sites a good firewall and a second connection usually achieve most of the same resilience for far less money and complexity.

Will SD-WAN fix a slow connection?

No. SD-WAN chooses intelligently between the connections you already have. If the connectivity into a site is poor, that has to be addressed first, and any provider telling you otherwise is overselling.

Why does the copper switch-off matter to this decision?

The old copper network is switched off on the 31st of January 2027. Businesses replacing lines because of it are already making connectivity decisions, which makes it a sensible moment to look at the wider network rather than replacing like for like.

Not sure whether SD-WAN is the right answer for your network?

We will look at what you run today and tell you straight, including if the answer is that you do not need it.

Talk to us