It is widely understood that first-wave SD-WAN was conceived as an OTT solution, driven by two fundamental assumptions: the underlying network is both insecure and unreliable. Based on these assumptions, SD-WAN solutions have developed their functions to frequently assess network quality and define ultra-responsive solutions in instances of quality degradation.
These assumptions, and the functions created in response have, however, created a mode of SD-WAN that is problematic and unsustainable.
While these functions may benefit some specific use cases, the cost-efficiency balance of these additional quality monitoring functions is skewed for most. Especially in instances where a customer is using MPLS, a network secured and reliable by nature.
A more balanced, mature approach is needed to enable the wider market to make the business case for, and truly reap the benefits of, SD-WAN. Currently, there are three major challenges SD-WAN is creating that are applying pressure to the network infrastructure: network traffic monitoring, outbound measurement and active probing, and encryption by default.
The top network pressures created by SD-WAN
Network traffic monitoring
As an unwelcome side effect of network traffic monitoring introduced by SD-WAN, networks often suffer from packet duplication to clear up and improve the quality of the traffic sent down the network. However, in turn, this inadvertently consumes more capacity…
Outbound measurement and active probing
Another challenging function is outbound measurement, a system that performs active probing and generates useless traffic when none is available. The most capacity-hungry solutions can consume up to 50Gb over two weeks in a single WAN interface.
Encryption by default
There is also a strong recurrence of ‘encryption by-default’, where areas of the network are overlaid and encrypted without assessment of importance or critical need. As a result, even on a secured and reliable network such as MPLS, the overall frame size and once again, bandwidth required, is needlessly increased.
Too often, these solutions are too sophisticated for their own good. And as a result, they are putting significant bandwidth pressure on the network.
A chicken and egg situation
SD-WAN is set to keep growing – with Gartner estimating 60% adoption by 2024 – but at this rate, we’re facing another network capacity issue. In its current incarnation, SD-WAN could end up requiring two or three times the bandwidth of normal traffic, becoming a primary contributor to poor network performance, that would then require yet more SD-WAN to compensate…
Many will continue to point to underlying network insecurities to justify their position but in truth, there needs to be a more critical assessment of what lies ahead. Irrespective of OTT supplier profits, this approach is unsustainable and irresponsible for building a futureproof network. Especially considering the world we find ourselves in today, with network demands higher and more diverse than ever before.
In fact, the Covid-19 crisis and the response of service providers has strongly demonstrated just how reliable and trustworthy underlying infrastructures are.
Why champion CSP infrastructure for SD-WAN
Recognising the strength of this underlying infrastructure sits at the core of this balance. And it is this that will also be crucial in making SD-WAN finally accessible to smaller players such as enterprises and SMEs who, up until now, have struggled to make the business case.
While ‘off the shelf’ SD-WAN solutions have commonly required additional hardware, new solutions are emerging that build SD-WAN as an extension of the existing hardware and core routing functions.
Now more than ever, networks are critical to business continuation. And to many enterprises and SMBs, SD-WAN can be crucial to enabling the agility needed to change their network behaviours and adapt quickly in adverse situations.
