SD-WAN: The underlay conundrum
Simon Clayton, Solution Architect at O2, considers what needs to happen for SD-WAN to deliver on the promises made of it.
Last month I wrote about the business case for a software-defined wide area network (SD-WAN), and some ways in which organisations could benefit from a managed solution. In this post, I want to focus on what SD-WAN can and can’t deliver, as well as alternative options you might consider.
When you read articles about SD-WAN, you’d be forgiven for thinking that it’s a cure for all. The technology that will solve any network problems that you currently have. On a recent bid I worked on, for example, seven of the eight submissions were for SD-WAN solutions, while only one was for a traditional, MPLS solution. Increasingly, it’s the solution offered, regardless of what the customer actually needs to do or achieve. In short, SD-WAN is the answer – now, what’s the question?
I want to make it clear that I absolutely love SD-WAN – it is certainly here to stay. Implemented properly, organisations will benefit from being able to leverage any combination of transport services to connect users to applications securely and dynamically, resulting in a more efficient network, with improved connectivity.
In my opinion, however, the problem lies in the promises made of it – and in the way that it is often referred to as a ‘virtualised network overlay’. The question to ask is what are you laying the SD-WAN on top of? What does your underlay network look like, and is it up to the job?
Service providers appear to be offering customers an SD-WAN based on internet access technology, whether access is DSL-based or wireless (4G/5G) based, or any other type of access for that matter. But whatever benefits it provides, ‘fixing’ internet-based access services for a customer is not one of them.
The starting point should be to look at your underlay network. And how you build this depends entirely on your business needs. For me, the questions to ask are exactly the same, whether you are planning an SD-WAN solution or not:
- What applications do your users need access to?
- Where are your applications based?
- Where are your users based?
- Do you need, or have private connectivity to the cloud?
- Do you have data centres?
- Do you need your connectivity to be guaranteed at all times?
- How sensitive are your applications to latency? Jitter? Loss?
- Do you want a contracted SLA for the availability of network resources for your users?
I recommend that you make sure these questions are answered before considering whether or not it will deliver the effective solution for you. First and foremost, it’s about building the right underlay network that will accommodate your systems and intended applications.
Whatever the claims, remember that SD-WAN doesn’t improve the availability of any network resource you require access to. Rather, it improves the efficiency of traffic to use network resources based on policies, defined by the IT admin. If you have a very low-speed ADSL, with poor transmission issues, or a 4G signal that struggles because your router is in the heart of your building, then installing an SD-WAN won’t make things better.
So, think about your underlay network before considering SD-WAN. The real USP lies in improving application performance – and this isn’t going to happen without a rock-solid foundation. It’s an exciting technology – but only if you build the right network, first time, to accommodate it.
If you’re ready to consider the additional automation and performance that SD-WAN delivers then we’re very keen to talk with you. We have published a white paper that will help you, which you can download here:
- SD-WAN: How to assess whether it’s right for your business
- Read the Gartner research about Managed SD-WAN Services
Check out our new research: Insights – How business and the workplace is changing.
Share your opinions on SD-WANs using the hashtag #O2Opinions.