IoT Architecture and Technology Protocols

Introduction
The Internet of Things (IoT) paradigm refers to a system of devices, interconnected with each other, equipped with computational capacity (smart objects), identifiable and enabled to transfer data over a network, without a required human interaction . The concept behind this paradigm is the pervasive presence of smart devices, which by cooperating with each other and interacting with human beings achieve common goals. The actual birth of the IoT dates back, according to Cisco estimates, to a period between 2008–2009, when for the first time, the number of connected objects exceeded the world population. In 2010, the number of such objects had almost doubled compared to that time, reaching about 12.5 billion. Since those years, IoT, thanks to continuous technological developments and considerable investments by companies, has become increasingly widespread in everyday life.
According to IoT-analytics estimates, there are currently about 20 billion connected objects globally, and the IoT sector generates a market of about $150 billion. In 2024, connected objects will exceed 30 billion, and the market value will be about 1 billion. As with any new technology trend, there are three possible categories of challenges for IoT to overcome: business, society, and technology.
The business field’s challenges mainly concern the identification of the motivation to start investing or not in a specific product and the design of a business model to achieve economic gain. In this category, depending on the use and the type of customer, products can be divided into three categories:
1. Consumer IoT (smartphones, smart car, smartwatch, etc.);
2. Commercial IoT (IoT Healthcare, Smart City, etc.);
3. Industrial IoT (includes various types of devices for industrial use).
The challenges in society’s field are to identify with the perspective of the customer who benefits from a product. To do this, it is necessary to consider some elements such as the constant change of requirements and demands imposed by the customer, the emergence of new devices, customer confidence in specific brands and products, and lack of knowledge of best practices in terms of privacy and security. Although the current technologies that belong to the IoT domain can now be defined as advanced, several areas can be identified that need further development.
IoT needs minimal components to be integrated into everyday objects. The miniaturization and integration of components itself is a field that can expand with the integration of silicon components into metallic or fabric materials. In addition, there is a need for such components to quickly harvest the necessary energy from their surroundings and use it profitably. Smart objects need to withstand harsh conditions, be it humidity, temperature, or shock and vibration; for their everyday use, they also need to be extremely reliable, and guarantee very high and consistent quality. Another aspect that is often underestimated is the ability of smart devices to self-configure and organize themselves. Moreover, it will be necessary to find standard protocols to identify objects uniquely. Moreover, a critical field concerns security to find solutions to secure connected objects, preventing cyber-attacks that can undermine the global growth of the Internet of Things.
Digging into the IoT Ecosystem
From a high-level perspective, the IoT landscape is a heterogeneous network, with the cloud computing layer being responsible for retrieving and acting on information gained from other layers.
IoT-related data can be stored in multiple locations within an IoT network. Sensors, gateways, and local devices in the network and cloud-based systems can store differing amounts of data.
Cloud computing
The cloud computing layer plays a critical role in the ecosystem, storing a vast amount of information and making decisions based on that data. It enables effective integration of data from the solution components. Adding the cloud to the IoT can also add security, availability, scalability, and performance, because cloud storage/database providers embrace those capabilities in order to achieve industry success.
The IoT smart home, for example, relies heavily on the cloud for its computing capabilities. The smart home has a sensor and home connectivity layer that collects data from different nodes and then provides this information to cloud servers for decision making.
This is where edge computing can help. Edge computing occurs when data is processed and stored at its source on the sensor or gateway, and the network only leverages the cloud if additional processing is required. Some endpoint devices don’t always send data back to the cloud. Instead, they use edge computing to store and process the data at the source. This helps enable a more real time experience for the end user. It also helps maintain the security of the network.
Edge computing can be especially useful if there are constraints on power or bandwidth. Using sensors to perform useful processing on data streams at the edge of the network reduces power consumption and uses bandwidth efficiently. Edge computing also helps protect user privacy by storing and analyzing data at the source rather than sending identifiable information to the cloud. As IoT technology advances and latency becomes a bigger issue, edge computing will become more widespread to enable real-time processing.
Most Common IoT Architectures
One of the main challenges to deal with the technological field to promote the deployment of IoT systems is to define a reference architecture that supports current features and future extensions. For this reason, such architecture must be:
• scalable, in order to manage the increasing number of devices and services without degrading their performance;
• interoperable, so that devices from different vendors can cooperate to achieve common goals;
• distributive, to allow to create of a distributed environment in which, after being collected from different sources, data are processed by different entities in a distributed way;
• able to operate with few resources, since objects generally have little computing power;
• Secure so as not to allow unauthorized access.
Currently, there is no single reference architecture, and creating one is proving very complicated despite many standardization efforts. The main problem lies in the natural fragmentation of possible applications, each of which depends on many very often different variables and design specifications. This problem must be added to each supplier’s tendency to propose its platform for similar applications. It is possible to see some of most common IoT Architecture used.

Three-Level Architecture
A generic high-level architecture composed of three layers has been introduced in the literature:
• Perception, which represents the physical layer of objects and groups all the features;
• Network, which represents the communication layer responsible for the transmission of data to the application layer through various technologies and protocols;
• Application, which represents the application layer in which the software offering a specific service, is actually implemented.
Perception Layer
The perception layer represents the physical level of objects and interacts with the surrounding environment by collecting and processing information. This level includes objects that, being able to interact with the external world and being equipped with computing capabilities, become in a certain sense “intelligent” or “smart”, where smart refers to the technological aspects (the smart technologies used), while intelligent refers to the functional aspects (self-identification, self-diagnosis, self-testing, etc.) of the sensor .
These smart objects, which are the fundamental blocks on which the IoT is based, can be objects of common usage (a refrigerator, a television, a car, etc.) or simple devices equipped with sensors and computing capabilities. In general, smart objects are equipped with the following essential properties:
• Communication: objects can connect to each other and to resources on the Internet to use data and services, update their status, and cooperate to achieve common goals;
• Identification: objects must be uniquely identified.
Depending on the specific application, one or more of the following properties may also be added:
• Addressability: objects can be directly reachable, i.e., addressed, to be interrogated and/or configured remotely;
• Sensing and actuation: objects can collect information about the surrounding world and manipulate it through the use of sensors and actuators; Embedded information processing: the smart objects are equipped with calculation
• capabilities to process the results of the sensors and drive the actuators;
• Localization: objects are aware of their physical location or can be located;
• User interface: objects can communicate appropriately with users via displays or other interfaces.
Examining the IoT Architecture
Some IoT applications have already come to fruition, such as the ability to wirelessly control a home thermostat or use a phone to open car doors. But the potential future applications of the IoT are much broader and larger in scale. Imagine using your phone to find a parking space or owning a refrigerator that knows when items need to be replenished and orders them automatically. Imagine sensors in bridges and other infrastructure that automatically inform engineers when repair or maintenance is required. Imagine location-based technology that informs you when you’re in close proximity to someone with a highly contagious illness or to a harmful substance or environment. These future applications will require IoT architecture platforms that perform a massive amount of data transfer and processing behind the scenes.
This IoT platform/architecture consists of several internally connected layers. Here’s a closer look at the individual components that exist within the IoT platform layers:
• Sensors and actuators: IoT sensors and actuators measure things like temperature, sound, moisture, and vibration. In a typical IoT smart home, a smart device like a thermostat has an embedded communication unit that connects to the home network. The sensors and actuators in the thermostat convert these physical measurements into electrical values that drive the system.
• IoT gateways: The gateway carries data between the local network and the Internet. The electronic values from sensors and actuators are received and then uploaded into the local network using network protocols such as Bluetooth, Bluetooth Low Energy (BLE), Cellular, LoRaWAN, Thread, Wi-Fi, or ZigBee. The gateway creates a meshed backbone to distribute the collected data and send responses to devices.
• Cloud-based IoT platform: The data transmitted through the gateway is stored and processed within a cloud-based IoT platform or in a company’s data centers. This data is then used to perform intelligent actions and make decisions.
• Applications: Ultimately, the data from IoT devices is used in applications to help people or organizations make better decisions or take specific actions. The applications push information from the cloud into applications on smartphones, tablets, or computers. The application layer is the most important to users because it’s their interface to the IoT network, allowing them to control and monitor the many elements of the IoT system, sometimes in real time.
Looking at Some IoT Protocols
Communication protocols form the backbone of IoT systems, connecting IoT devices to the network and ultimately to applications and users. These standards-based and proprietary protocols enable data to pass between the different layers of the IoT architecture by defining data exchange formats, data encoding, addressing schemes for devices, and the way that data packets are routed from node to destination.
The IoT ecosystem includes a range of different protocols supporting short-range, local, and wide area networks, all of which coexist. Each technology has specific characteristics in terms of range, sensing and control, and the ability to transmit different types of information. Together, these technologies can provide seamless coverage at all wireless ranges and capabilities. For example, Bluetooth works well for short-range applications, while narrowband IoT (NB-IoT) works well for long-range applications.
Here is a summary of some key network protocols:
• The IEEE 802.15.4 standard is a collection of Low-Rate Wireless Personal Area Network (LR-WPAN) standards. These standards provide low-cost, low-speed communications for power-constrained devices. They form the basis of specifications for high-level communication protocols such as Zigbee. Zigbee is a mesh network designed for low-power operation, used in smart homes and in smart-energy applications for utilities. Zigbee is based on the IEEE 802.15.4 Physical Layer (PHY) and Medium Access Control (MAC) standards.
• Wi-Fi is a collection of IEEE 802.11 Wireless Local Area Network (WLAN) communication standards. Wi-Fi provides high data rates to both indoor and outdoor locations and is very widely used.
• Bluetooth is an open standard maintained by the Bluetooth Special Interest Group (SIG). It is a low-cost wireless communication technology suitable for data transmission between mobile devices over short distances, such as 8 to 10 meters. It’s used in applications like audio streaming, cars, speakers, and headsets.
• Bluetooth Low Energy (BLE) is part of the Bluetooth standard. It’s designed specifically for lower-power operation. BLE devices commonly use coin cell batteries to operate. BLE is used in IoT devices such as light bulbs and light switches.
• Thread is a low-power, secure, and Internet-based mesh networking technology for IoT products. In 2014, the Thread Group was formed as a working group to help the adoption of Thread. Thread supports existing IPv6-based connectivity standards, on its secure, low-power, mesh network.
• LoRa is for long-range communications and is a low-power wide-area network (LPWAN) protocol developed by LoRa Alliance. This technology is optimal for enabling sensors for large-scale agriculture applications.
• Cellular standards like 5G provide a network backbone for IoT services, supporting both high data rates and long-range communications. A commonly used cellular IoT standard is NB-IoT, used in smart parking, utility management, and manufacturing automation.
These messaging protocols are used to share data across devices and with the cloud. The IoT protocols are a critical part of the IoT technology stack and, without them, hardware simply wouldn’t work. IoT protocols enable the IoT device to exchange data in a controlled and meaningful way.
For example, smart devices of the IoT are called “smart” because not only are they able to talk with each other, but if they encounter problems they can automatically mitigate the issue or call for help via the network. This interaction is only possible through protocol communications or a common language that the IoT devices are given.