Edge computing could fail, if you only pay attention to the tail

In its obsession with software development, all too often at the expense of other disciplines, the IT industry has let the tail wag the proverbial dog for more than ten years. Now, with the rise of edge computing – where your architectural and operational decisions will be absolutely key to your success – it risks losing the dog altogether.

It’s true that Steve Ballmer’s chant on stage of “Developers, Developers, Developers, Developers,” back in 2008 at Microsoft’s 25th anniversary event, made absolute sense at the time. His performance became a huge Internet meme that sparked IT vendors across the board into declaring deep love for the software development community.

However, it’s brought us to a present-day where the typical DevOps discussion places a huge emphasis on the ‘Dev’ piece and relatively little on ‘Ops’. Yet when it comes to the bottom line, it’s Ops that generates the revenues, while Dev is in many ways just a cost!

Fortunately, we do seem to be regaining some balance, with the disciplines of architecture and operations coming back in focus across the industry. It’s now pretty well accepted that success with digital transformation is about much more than the development of good-looking web and mobile apps.

Systems, old and new, need to be properly integrated and work end-to-end in a secure, robust and efficient manner. Customer satisfaction can be hit hard if the back-end doesn’t meet expectations created by the front-end. Disjointed systems also constrain how fast you can respond to new and changing needs and opportunities.

Ops matters more than ever, as computing moves to the edge

All of this means that Ops really matters, and this is even more true as we move into edge computing. This is an area in which you’re going to struggle or fail if you take a developer-led approach, treating the platform, integration and operational considerations as secondary discussions that can take place downstream.

The truth is that the most volatile part of the edge computing equation is what takes place at the edge itself. Devices, sensors, machine learning models and other software components will come and go over time. Making architectural decisions based on the current mix of technology and applications at the edge is simply not sustainable.

In order to implement a relatively future-proof edge computing environment, you have to start with sound architectural design. This not only means thinking through the dependencies and flows across the various tiers from the edge to the cloud, but also considering how provisioning, configuration, monitoring and administration of systems and components will be handled on an ongoing basis. The phrase ‘design for operations’ sums up the mindset to adopt here.

Building on this concept, our own observations of edge computing and industrial IoT initiatives confirm that the need for end-to-end architectural and operational thinking goes way beyond of the IT domain. In an Industrial Digital Transformation context, for example, IT professionals typically need to work with engineers specialising in plant, machinery and other edge resident equipment, whether that edge is a factory, office building, power station, oil pipeline or smart vehicle.

Against this background, the most critical question you need to address before starting any edge computing initiative is who needs to be involved.

Data Center
Data Management