The rise of cloud computing has been one of the defining trends of the past ten years – and it is projected to keep growing rapidly in the next ten.

But in the meantime, the explosion of the Internet of Things (IoT) will redefine the way data is created and processed – so much so that, by 2025, 80% of our data will actually be processed outside of the cloud and data centers, compared to 20% today.

Where will it go? To the Edge !

To understand this, we need to get back in time a little.

In the past ten years, cloud computing has taken the IT world by storm. It has become a part of our daily lives, both personal and professional.

Recent surveys show that today, 19 out of every 20 companies use the cloud and data centers. And the billions of people who talk with their friends on Whatsapp, listen to their favorite music on Spotify or check their emails on the go use the cloud daily, sometimes without realizing it.

It may therefore seem a bit weird or provocative to announce the dawning of a “post-cloud era”.

For, in the foreseeable future, the cloud and data centers are here to stay… And they will actually keep growing rapidly to accommodate ever increasing needs.

But, when we look at the big picture, the volumes of data that will be processed outside the cloud will grow even faster.

 

The rise of the machines

To understand why, we first need to remember that, by 2025, the vast majority of the volumes of data we generate will not be created by humans, but by objects and machines – What we call the Internet of Things.

By then, these objects – the smart sensors that equip our cities, the increasingly connected vehicles in our streets, the sensors and machines that operate in warehouses and factories, among many others – will account for nearly 90% of the volumes of data in the world.

And for these objects and machines, sending everything to the cloud is not the best option. In some cases, it’s not an option at all.

 

The end of “Cloud by default”

One of the issues is the sheer volume of data we are talking about.

By 2025, there will be more than 80 billion of connected devices in activity – imagine for a moment the storage and network bandwidth that would be required to send all their data to the cloud !

Another problem is latency and response times.

For some uses, some latency is alright: when I am to listening to a podcast, it may indeed be annoying, but not disastrous.

By contrast, for an autonomous vehicle, latency or lack of connectivity may literally be a matter of life and death – the difference between braking at the right time and braking too late, for example.

And, if the example of the car is undoubtedly the most telling, it is only one of the many uses for which latency times and network availability are a factor.

In these situations, we need to process data much closer to the source, at the periphery (or edge) of the network – and this is what Edge computing is all about.

 

Cloud and Edge computing, two complementary technologies

For these cases, data need to be processed by the device itself (directly in the car, in the phone etc.), on the premises, by cell towers or small local data centers – in other words, at the edge of the network.

The cloud and data centers will continue to be the core of the network – but not all data will go through them.

For companies, Edge and cloud can offer a great combination, allowing them to choose which data will be processed locally and which will be stored in the cloud. It will mean greater control over their critical information.

Edge infrastructure will also make it possible to bypass the cloud when it’s not available, switching from one option to another to avoid any delay or loss of information.

For some IT professionals, this may be familiar territory: Atos, for one, has been a pioneer to implemented on-premise AI, to adapt to physical constraints (eg in the mining industry) or to handle strategic data.

But for many businesses, it will be a significant change in the way their data is handled and in the way they orchestrate their infrastructure.

 

Bringing AI closer to the source of data

This change will also bring many opportunities and possible benefits.

By 2025, 30% of the data will be real-time data. Edge computing will be used to manage and analyze it much faster.

It will also make it possible to run artificial intelligence locally on embedded systems, which can be just as useful for traffic management as it is for robots in a production line.

In a recent use case, we made it it possible for wind turbines to communicate with one another, only sending necessary information to the Cloud, and to adapt to network connectivity problems.

In the field of medicine, Edge computing will be indispensible to create “robot surgeons” that can adapt in real-time to what they see and to the patient’s situation. In this case, it is better not to rely on a 4G or Wi-Fi connection!

In construction, this will be how intelligent buildings will run routine programs to reduce their energy consumption through connected lights and thermostats.

 

The future is near

And this processing of data “close to the source” can also have many positive side effects, for businesses as well as consumers.

It can limit the quantity of personal and private data that goes to the cloud. At a time when many raise concerns regarding the storage of data such as voice commands, facial recognition or fingerprints, Edge computing can provide a more secure and privacy-friendly solution.

This subsidiarity allows for greater data security. This point is critical, as cybersecurity is still widely viewed as the Achilles’ heel of IoT.

As such, Edge computing can bring answers to some of the most pressing IT conundrums.

This is why it will be the main topic of my keynote at Vivatech on May 16 and the focus of the Atos Tech Days, held in Paris on the 16th and 17th of May. They will include real-world use cases… and several world-premieres!

So – are you ready for the Edge revolution?

 

Click here to visualize the article