The Internet of Things: toward a vast distributed computing system

For a long time, the Internet of Things was relatively simple to picture. A sensor measures a temperature, a camera transmits an image, a vehicle reports its position, or an industrial machine provides some data about its operation. That information then travels over a network to a server or cloud platform where it is stored, analyzed, and presented to users.

This representation remains valid for many use cases, but it describes less and less well what the IoT is becoming.

Connected equipment is progressively gaining computing power. It has a digital identity, uses several types of networks, communicates with different platforms, receives remote updates, and can execute certain functions locally. The infrastructures surrounding it are evolving as well: they manage connectivity, authentication, routing, security, data, and integration with the cloud. Edge computing brings processing closer to where data is produced, and artificial intelligence now makes it possible to interpret certain situations directly in the vicinity of the physical world.

The Internet of Things is thus evolving toward something far broader than a collection of connected sensors. It is progressively becoming a vast computing infrastructure distributed across the cloud, networks, the Edge, and millions of devices present in the physical world.

This transformation profoundly changes how organizations must design, secure, administer, and govern their digital infrastructures.

Enterprise computing was long relatively centralized. Data, applications, and computing power were concentrated in a few clearly identifiable systems and infrastructures. The Web, mobile devices, SaaS services, and the cloud progressively distributed this environment. The Internet of Things now pushes that logic much further by moving part of the computing infrastructure directly into the physical world.

A manufacturing company may operate connected equipment across several plants. A transportation company may manage thousands of vehicles. A municipality may deploy sensors throughout its territory. An energy producer may own equipment located in remote regions, and a manufacturer may sell devices that will continue to communicate with its systems for ten or fifteen years.

Part of the computing infrastructure is then literally located in the plants, vehicles, buildings, equipment, and products used by customers. These resources may be spread across several countries, use different networks, and remain physically inaccessible for long periods. The traditional notion of an IT perimeter becomes far less representative of operational reality.

The word “thing” itself can become misleading. A modern connected device may have a processor, memory, an operating system or firmware, certificates, cryptographic keys, a digital identity, network interfaces, applications, and update mechanisms. In many respects, it looks more like a specialized computer than a simple peripheral.

Edge computing accentuates this evolution. Some data can now be processed directly in the device or in an infrastructure located nearby. A camera can perform an initial image analysis before transmitting a result. An industrial machine can detect an anomaly locally. A vehicle can make certain decisions without waiting for a response from a data center hundreds of kilometers away.

The device then stops being only a source of data. It becomes an active participant in the computing system.

This transformation does not, however, diminish the importance of the cloud. It creates instead a complementary architecture. The cloud remains particularly effective for centralizing certain data, training artificial intelligence models, performing large-scale analytics, coordinating device fleets, and delivering global services. The Edge brings proximity to the physical world.

That proximity makes it possible to reduce latency, maintain certain operations when connectivity with the cloud is temporarily interrupted, and limit the volumes of data to be transported. In some contexts, it also makes it possible to process sensitive information locally and transmit only the necessary events or results.

The architecture therefore becomes distributed out of operational necessity. Some functions belong to the device, others to the Edge, the network, or the cloud. The architectural question then becomes: where must each capability reside in order to obtain the right combination of performance, security, resilience, cost, and sovereignty?

The network itself evolves within this architecture. In a traditional conception of the IoT, it could be seen mainly as the means of transporting data between the object and a central platform. As infrastructures become more complex, the connectivity layer can also take part in identity management, communication control, device segmentation, data routing, and the enforcement of certain security policies.

The network thus becomes more than a pipe. It takes part in orchestrating the infrastructure.

This evolution becomes particularly important when an organization operates thousands or tens of thousands of devices. Administering a few sensors individually remains relatively simple. Administering a fleet of one hundred thousand devices spread across several regions completely transforms the nature of the problem.

Each device has a state, a software version, authorizations, communication destinations, and a configuration. It produces data, may require an update, may fail, may be compromised, may change owners, or may need to be removed from the network. The fleet then becomes a living system that must be inventoried, monitored, updated, and secured throughout its life cycle.

The ability to manage the life cycle becomes as important as the ability to connect the device in the first place.

This reality places identity at the center of the architecture. In an infrastructure made up of thousands or millions of devices, knowing who is communicating becomes as important as knowing what data is being transmitted. The system must be able to recognize a legitimate device, determine the resources it can use, and control the actions it is authorized to perform.

Identity management therefore no longer concerns mainly humans. Devices, applications, APIs, cloud services, and AI agents now join employees. Each potentially represents a non-human identity with its own permissions and trust relationships.

This proliferation profoundly transforms IAM and Zero Trust. A connected device should not obtain permanent trust simply because it was installed by the organization. Its identity, its state, its context, its permissions, and its behavior can evolve over its lifetime. Trust must therefore be verifiable and subject to reassessment.

This is where the notion of Continuous Trust becomes particularly relevant. In a distributed infrastructure, trust can no longer be treated as a state acquired once and for all. It must become a property continually evaluated based on identity, behavior, context, and risk level.

Artificial intelligence now adds another dimension to this infrastructure. Specialized models can be run locally on devices or Edge infrastructures. A camera can identify a specific event, an industrial machine can recognize abnormal behavior, a logistics system can detect a situation requiring intervention, or a device can analyze certain data before transmitting it.

The distance between observation and decision then shrinks considerably.

This capability also transforms how data flows. Instead of continuously transmitting large volumes of information to the cloud, an infrastructure can perform part of the processing locally and transmit only the relevant events or results. This approach can improve performance, reduce certain communication costs, and limit the movement of certain sensitive information.

It brings with it, however, a major consequence: intelligence itself becomes distributed.

The organization must now manage devices, data, and communications, but also models, versions, permissions, and decisions executed in different places across its infrastructure. Questions of artificial intelligence governance then progressively leave the data center to join the equipment present in the physical world.

When intelligence is distributed, so is risk. A server installed in a data center generally operates in a controlled environment. An IoT device may be located in a plant, a vehicle, a store, a public building, or directly at a customer's premises. Some devices may be physically accessible, have limited computing resources, and remain in service far longer than traditional applications.

Cybersecurity must therefore simultaneously cover the device's identity, its operating system, its communications, its update mechanisms, its permissions, the data it produces, and the infrastructures it interacts with. A vulnerability located in just one of these layers can affect the whole.

The principle of least privilege becomes particularly important. A sensor responsible for transmitting a temperature has no reason to communicate with the company's entire network. A camera should not access every resource it is technically capable of reaching. The more distributed the infrastructure becomes, the more precisely trust relationships must be defined and controlled.

This reality perfectly illustrates the evolution toward Hypersecurity. Protecting a distributed infrastructure no longer consists solely in securing devices, networks, or the cloud individually. It requires understanding the relationships among identities, data, applications, communications, AI models, and physical infrastructures. Security lies as much in the interactions between these components as in each of them.

The shift becomes even more significant when the IoT meets artificial intelligence agents. A sensor can observe a situation, a model can interpret it, and an agent can then consult other information and determine that an intervention is necessary. That agent can then trigger an action in another system.

The loop between observation, interpretation, decision, and action can then progressively close.

This capability opens up considerable possibilities in industry, logistics, energy, transportation, agriculture, smart buildings, and many other sectors. It simultaneously increases the importance of governance. The faster and more automatically a system can act, the more precisely its authority, its permissions, its limits, and its escalation mechanisms must be defined.

The distributed infrastructure thus progressively becomes decision-making.

This transformation makes Edge architecture a strategic decision. Choosing where data, identity, security controls, intelligence, and the capacity to act reside directly influences performance, costs, resilience, confidentiality, and sovereignty.

An excessively centralized architecture can become heavily dependent on connectivity and require transporting considerable volumes of data. An overly distributed architecture can become difficult to administer and secure. The challenge therefore consists in intelligently distributing capabilities according to operational needs and risks.

This distribution must also be able to evolve. Networks will change, artificial intelligence models will advance, new vulnerabilities will be discovered, and regulatory requirements will evolve. A device installed today may still be in service when several of the technologies used when it was designed have disappeared.

This reality requires thinking about the IoT in terms of life cycle rather than initial deployment. Connecting a device and receiving its first data is only the beginning. True success consists in being able to administer that device for five, ten, or fifteen years. Its software will have to be updated, certain certificates renewed, its permissions modified, and its communications adapted. The services it connects to may themselves change.

An architecture that works perfectly on the day it is deployed can become extremely costly if every change requires physical intervention. Remote administration capability, renewal of identities and cryptographic mechanisms, modularity, interoperability, and the ability to evolve certain components therefore become essential characteristics.

This perspective ultimately brings the IoT back to a question far broader than connectivity: how do you sustainably administer a computing system a significant part of which is scattered throughout the physical world?

The Internet of Things is progressively moving beyond the model of a sensor transmitting information to the cloud. Equipment is becoming computing nodes, the network plays a greater role in the architecture, the Edge brings processing closer to the physical world, and artificial intelligence makes it possible to interpret certain situations directly where they occur. With AI agents, this infrastructure can progressively move from observation to action.

We are thus seeing the emergence of a vast distributed computing system whose components may be found in data centers, in the cloud, on telecommunications networks, in plants, vehicles, buildings, or millions of devices deployed in the field.

For organizations, the challenge therefore goes far beyond the ability to connect this equipment. They must master its identity, its life cycle, its access, its communications, its updates, the data it produces, and the intelligence models that may run on it. They must also determine where processing and decision-making capability should reside in order to maintain the appropriate level of performance, security, resilience, and sovereignty.

This evolution connects directly with Quantum Beyond's approach to Hypersecurity and distributed intelligent infrastructures. Qb OS Edge, Q-Carbon Security Systems, IAM, Zero Trust, and Continuous Trust can help build environments where equipment, identities, data, and new intelligence capabilities remain governable throughout their life cycle. Our experts work alongside technology and operations teams to integrate these different dimensions into architectures capable of evolving along with uses and risks.

For a long time, we connected objects to our computing systems. A deeper transformation is now under way: the objects are themselves becoming part of the computing system.

When computation, identity, intelligence, and the capacity to act are distributed all the way into the physical world, the Edge stops being a mere periphery of the digital infrastructure. It becomes one of the places where the digital enterprise exists, decides, and acts.