The history of packet data has a long list of workarounds to address our problems of quality and performance. That’s not sustainable, so will change.
You are already under pressure to deliver today’s Web and Unified Communications with a better experience and at a lower cost. You know that you soon will be asked to support virtual reality, automated homes, smart contacts, and revolutions in energy, health, transport and more.
You also know that optical fibre and 5G technologies, whilst valuable, are not on their own going to solve your application service quality problems. After a quarter of a century of trying, it has become obvious to everyone that more quantity (i.e. “speed”) has been exhausted as a “cure all” for broadband quality problems.
So what can you do? You have a strategic choice.
One option is to adopt yet more technical workarounds: even greater over-provisioning, increased traffic management, smarter data caps, deeper CDNs, enterprise SD-WAN, machine learning for network configuration, and more adaptive applications. These techniques have been refined over time through trial and error, as we uncover the past mistakes of our architectures, and limits of our measures and methods.
This option will, to a degree, help you to improve quality. However, it will never deliver assured application performance. You will forever be stuck in the present commodity broadband Internet access model. Your last resort will be a commercial workaround: to over-promise, under-deliver, and treat the SLA breach as a dishonourable cost of sale.
The other option is to address the root issue: the missing science and tools for network performance engineering. That requires us to overcome our historic limitation: to see latency through the user’s eyes, break it down to its constituent causal parts, and take control over its allocation. With engineering, your networking products and services will say what they do, and then do what they say, with a predictable safety margin.
The benefits of workarounds are falling due to diminishing returns, and the costs are rising due to compounding complexity. Ultimately, the workarounds go into reverse, driving systemic failure through non-stationarity.
Meanwhile, the technical and commercial barriers to engineered quality are falling fast: we now know how to make, sell, and buy it. There are working examples of the network mechanisms; the engineering methods and new mathematics are proven; and the management science has decades of application in other industries.
The timing is right for a paradigm change.
The new burning telecoms and cloud industry issue is the innovative adoption of quality management, not its invention. I believe the right choice for most people is to “take the high road”, and to start using network performance science to engineer assured cloud application performance outcomes.
This ability to precision engineer quality will enable you to sell new products that cannot be replicated by competitors using complex workarounds. It puts yourself and your organisation on the next ‘S-curve’ of growth. The journey takes us to a place of honour: an engineered infrastructure for our grandchildren to inherit.
It may take courage to be an engineering pioneer, but the endless workaround alternative is a painful road. After all, our homes no longer burn down just because we seek a small light at night. Maybe our networked applications should stop accidentally combusting, too?
For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.
Read It Now
