With the rise of connected objects, IT professionals and users alike are witnessing a transformation in the use of the internet and its ecosystems.
We are connected to them through IoTs. They are rapidly becoming part of our pockets, our wrists, our furniture, integrating us almost completely into the digital sphere.
So it’s interesting to know how this works, what the infrastructure and data issues are, but also the impact on business opportunities.
The Internet of Things or IoT is the interconnection between the internet and physical objects such as :
- SmartWatches and home automation objects
- Autonomous vehicles
- Medical sensors
And many others, including space exploration.
Considered as the third Internet revolution, this constantly evolving field is constantly growing thanks to innovations that are rapidly integrating the consumer market.
Thus, by 2025, it is estimated* that nearly 150 billion objects will be interconnected worldwide.
All these objects will have to send, receive and process increasingly complex and dense flows of information, in an immediate manner.
With the industrialisation and democratisation of the IoT, the mass of data from the various sensors (audio, video, text, image, heart rate, traffic, etc.) must be processed with the least possible latency.
Moreover, given the diversity of these systems and the current lack of standardisation, the challenge of compatibility is also very high.
The responsiveness required of these devices requires a great deal of computing power, in addition to rapid access to reference sources.
This is where the power of the Cloud (Azure, AWS, Oracle, etc.) and Edge computing comes into play, enabling this level of responsiveness to be achieved.
Indeed, each IoT is able to manage/analyse local source data, while increasing its capacities through the network, in real time.
AI and machine learning for IoT relevance
These capacities for analysis, processing, interpretation of data, as well as relevance of restitution, require a certain amount of on-board computing power.
In this field, two approaches coexist:
- Embedding high computing power
- Embedded AI or machine learning.
While the first approach consists of embedding ultra-fast processors** that will collect information in real time and then process it according to a predefined algorithm, it is more quickly constrained by hardware and software limitations.
The second solution consists of integrating a previously “trained” AI model, not requiring large resources, which will compare its inputs with a large panel of hosted evolutionary models, allowing it to “learn” and improve as it is used (Alexa, Cortana, Google Assistant…).
This second option is chosen by most IoT manufacturers.
Who says IoT, IA, interconnection and reactivity, says network power and connectivity.
While fiber networks offer ultra-fast connections for local IoTs, “mobile” connectivity technologies offer, for certain applications, limiting speeds.
So 5G technologies seem to be promoted and indicated to meet the future needs of these technologies.
This network, in full deployment, will offer a denser and more efficient mesh than 4G, with higher theoretical speeds (+ 500Mb/s against a maximum of 500Mb/s). This should make it possible to support future technical developments, in particular concerning autonomous vehicles and IoT in medicine.
Indeed, the use of IoT poses several problems related to their use:
- Cybersécurité :
- Encryption of data, prevention against attacks and hijacking of IoT functions.
- Data confidentiality:
- Protection of privacy, anonymization, certification, exploitation.
- Storage:
- Sizing, deployment of cloud infrastructure, storage optimization.
- Development :
- Advancements in machine learning, acceleration of deep learning, source verification, blockchain.
This demand for IT skills is already underway and is resulting in the creation of attractive opportunities, as well as salaries that, for seniors, can reach more than 10K CHF/month***. Moreover, these new technologies can transform generalist IT professions and require more specialists in the long term.
* ETH Zurich ** Nvidia processors for autonomous vehicles *** Monthly salaries for senior IT professionals – ICT Journal.
