What CIOs need to know about Edge Computing

The role of CIO is constantly evolving, twisting and turning to match the rate of change in IT itself.

In recent years, we've had waves of technology change and shape our IT infrastructures and challenge the way we do business.

Take the Internet of Things, for example. This now reasonably well understood technology has influenced IT departments and business offerings. And Edge Computing is extending its evolution.

The term has exploded in recent years, with an increasing number of organisations researching its potential.

In its simplest form, Edge Computing is the method of optimising cloud computing systems and taking control of their applications and data.

Diving further, the 'network edge' refers to the extreme point of the network: the clients. Edge Computing refers to the work carried out on these devices.

Specifically, these days this usually concerns IoT devices that do not necessarily send sensor information to the cloud, but process the telemetry on the device itself.

That sounds like a small difference, but it has enormous network consequences.

In a bid to make better and faster decisions, NHS Blood and Transplant Chief Digital Officer Aaron Powell is exploring Edge Computing to extend his data analytics capabilities.

He said: "The interesting thing around Edge Computing is the emerging sense that we have often sought to take the data to the analytic capability.

"I think increasingly we are finding that's not the way that's going to work anymore, you'll find you need to take the analytic capability to where the data is.

"As both a way of reducing the cost of moving the data around which can be quite expensive in itself, but also enable us to apply the analytics even more quickly. Because if you can take that capability to where the data is you get quicker decisions.

"I think that capability of moving the analytics to where the data is is the emerging theme for us," the CIO 100 member concluded.

Away from the server

The idea seems to be the opposite of the thin client, in which almost all the calculations are carried out on central servers.

In the Cloud model, intelligent tasks are performed on servers and then passed on to 'stupid' devices. With Edge, that work shifts more to the clients.

However, Edge and Cloud are not opposing concepts. Edge Computing is not so much about intelligence on the device itself, but about working as close as possible to the source data.

This not only means that it offers options for data minimisation - raw source data doesn't have to be all shipped to the cloud - but also for very powerfully distributed systems.

Powerful Distribution

The dream for Edge Computing is that millions of IoT devices together form a massive intelligent network that performs tasks that can now only be seen in a huge data centre. And that is quite different from the cloud models of today.

In the most obvious network model today, you have a combination of Edge and Cloud: distributed systems process data on the devices themselves and transmit the results to the cloud where they are stored or perform analysis that is not (yet) possible within distributed systems.

Is this the end of the cloud?

Distributed systems are the future, simply because we are currently loading the world with networked devices.

Think, for example, of millions of thermostats that work together on climate models or networked cars that analyse traffic information themselves instead of drawing it from servers.

That results in a huge backhaul of data to central servers and for that reason alone it makes more sense to split up work across all those networked devices themselves.


Copyright © 2018 IDG Communications, Inc.

7 secrets of successful remote IT teams