SD-WAN, and Software-Defined Networking (SDN) more generally, is a game changer. The rules have been torn up. An SD-WAN looks nothing like a regular WAN.
This requires a change in mindset and that’s causing a problem for people across the industry.
Those running IT functions have less of a leap to make. They’re focussed on delivering value for their users and the wider business and will use any technical means to do so. Every survey of IT decision makers and managers shows a growing desire to explore what SD-WAN can do for them.
But service providers (ISPs – those with network infrastructure), are struggling – and here is why I think that is.
ISPs, in common with most industries, have a business investment cycle. Either they have a rolling programme of changes and upgrades to network equipment, and (if they are growing) bandwidth upgrades to cope with increase in demand, or every few years they invest millions in a complete overhaul of their equipment.
Sometimes this involves building a completely new network, in parallel to their existing one, and porting over customers individually or in large batches, and eventually decommissioning the aging, legacy network.
The business cycle results in disruption (when a migration doesn’t go quite according to plan, or when an upgrade fails), and in fluctuations in quality (great after the investment, less so when those assets are sweated and the Finance guys delay the next upgrade as long as possible).
But more importantly, it results in repetition – doing the same thing, over and over again.
It’s the least risky thing to do isn’t it? It worked last time, and the company is still here, so lets just do it again, but with better kit.
But this time, the change driven by SDN is so huge that none of the service providers are ready for it. And the bigger they are the more entrenched in traditional ways of networking, and investing in that network, they are.
In the past, upgrades in technology have been about how fast something is, or how much capacity it has. With every network upgrade has come a smaller footprint of hardware, with more capability.
That’s great, but never has it addressed a fundamental issue with networks, and that’s the potential to disaggregate the software from the hardware.
Ok, I can tell I’m starting to lose you now – but here’s why it’s such an important message, and one communication providers the world over are struggling with.
Up until now, the software and hardware come together. Think of buying an iPhone – It’s inconceivable you would separate the software and hardware – it all comes from Apple in one integrated package.
But SDN and SD-WAN is about separating those two elements, and to great effect.
Creating a level of abstraction between the hardware (which as a result can be cheap and cheerful, and unbranded) and the software (where all the control is) means you can create a network, and networking products, that can scale almost infinitely.
No ties to particular, expensive brands of hardware, or particular types of circuit, yet the power to change topology on a whim, and add resilience and QoS where they were difficult or non-existent before.
And so, service providers are reluctant to make that leap. To completely tear up their network, and their existing supplier relationships, and to start afresh. To invest in a network of cheap hardware, knowing they can run any software they like on it, and keep scaling up.
Which brings me to my final observation. The software. It’s not being written by the service providers.
And this is their greatest weakness. ISPs need to become software houses. They need to write their own networking software, not use other people’s. They need to build value, and not be tied to any third party. And they need to do it because no one else is going to bridge the gap between their customers and their network.
Providers that write their own software, rather that licencing it from abroad, are going to have a huge advantage over the competition in the coming decade and beyond.
Sure, some service providers will eventually take the leap and install generic hardware, and then run someone else’s software on it – and that will be a step in the right direction. But they will still have that disconnect, and that lack of power.
Until service providers write their own software for their own network, they will never quite realise the enormous benefits of SD-WAN.