Taking off with edge computing? Don’t clip your own wings

Posted by Marc-Elian Bégin on 7 February 2022

Considering an edge solution? Read on for tips and potential pitfalls.

A pilot entering cloudy, turbulent skies is effectively flying blind. It’s a critical situation. Every moment that ticks by without clear visibility means things can change and danger levels increase. It may become impossible to reach the destination airport. Data is needed immediately – facts about the aircraft, its status, position, nearby hazards, and weather conditions – so the pilot is able to proceed without guesswork and regain full control.


Through my aeroclub, I’m not limited to flying just one type of aircraft. I like to keep my options open.

As a private pilot, I know a lot about airplanes but I’d never trust myself to build one and take to the skies! A do-it-yourself (DIY) solution – even one made from excellent parts – carries too much risk. I feel the same is true for edge computing.

Too many times, I’ve seen DIY edge solutions fail. Companies realise too late the downsides of unbudgeted cost, protracted time to readiness and operational complexity. And they clip their own wings by sacrificing system reliability and evolvability.

I guess people are tempted to self-build because edge computing seems similar to cloud computing. Sure, they share several attributes but they’re also very different.

Take, for example, the physical environment. Cloud computing happens in highly controlled locations with easy-to-access servers. But the edge is challenged by temperature and humidity changes, dust and vibration. Unregulated settings get power and network jitters, have more cyber-security risk and can be complex to access.

Instead of a DIY edge solution, some organisations ask a vertical solution provider to solve a specific use case. The danger here is a significant investment with little flexibility in the future. As business needs evolve, the original solution may not adapt. A flexibility promise might be made but it will be on the provider’s terms, with limited powers of negotiation.

An obvious alternative is a cloud-first approach. Start with a cloud from one of the hyperscalers (Amazon, Microsoft Azure or Google) or a private cloud, and deploy at the edge as mandated. Warning! There is a conflict of interest. While you seek the best possible edge outcomes, these providers are in the business of selling compute and storage. They want you to pump as much data as possible to the cloud. So you’ll miss what’s arguably the No.1 benefit of edge computing: processing raw data near its source.

Cloud-first can work for simple deployments but – once again – you risk clipping your own wings. As your business needs grow in scale and sophistication (trust me, they will), a cloud approach quickly limits your ability to expand. At which point, you may regret putting all your eggs in the same cloud basket.

So the optimal approach is edge-first and cloud-neutral. This aims to privilege the edge (instead of the cloud) because this is where you act on your data. An edge-first solution does not limit your ability to process, transform, aggregate and anonymise the data. On the contrary, it generally empowers developers with a wide range of tools (including machine learning) to take full advantage of the data produced at the edge and act on it. And you do all this on your own terms while also leveraging the unlimited processing power and storage of the cloud. You get the best of both worlds – edge and cloud.

Another way to keep your options open, ensuring the most future-proof edge experience, is to select a provider offering OPEX over CAPEX. Added to this, edge-first providers are likely to help you deploy a shared infrastructure strategy. This allows several old and new apps from different sources to run alongside each other at the edge, which is likely to generate a very attractive return on investment.

One final thought. Data sovereignty matters. Keeping tight control of where your data gets processed is strategic for many organisations. A key decision is which parts of your data will be processed at the edge and which in the cloud. The solution you select must give you the right level of control over your data flows. 

An edge-first, cloud-neutral approach means you won’t clip your own wings. We will be happy to help you on your journey to the edge. So please get in touch for lock-in free, future-proofed advice.

Engage on LinkedIn

Ready to start your journey to the edge with our edge-to-cloud open platform?

Go to Nuvla.io