solutions have their pros and cons, as discussed in Section 1.2. Therefore, when choosing between licensed and unlicensed, it is important to pay attention to the application needs and the trade-offs between different technologies. In general, cellular IoT device’s cost and service fees tend to be higher than unlicensed alternatives. On the other hand, unlicensed options have limited capabilities concerning scalability, coverage, and security.
1.4 Key Enablers
In this section, we discuss some enablers for efficient connectivity in massive IoT deployments. First, we discuss a holistic and scalable system, which aims to address the networking and the potential of global scalability. Next, we focus on enablers for sustainable connectivity.
1.4.1 Holistic and Globally Scalable Massive IoT
Holistic Massive IoT. With the increasing demands of massive IoT applications, the network should support dynamic and collaborative orchestration between end-to-end applications, often communicating over multiple domains (wireless, wired, optical), involving multiple radio access technologies (RATs) operated by different stakeholders (public, private, cellular). Thus, the network handles the requirements, even though the application becomes agnostic of the interfaces [35]. For instance, [35] proposes a step toward such integration, since it envisions LPWANs complementing cellular services in many massive IoT applications. At the same time, LPWAN resorts to cellular IoT (and broadband) to extend and supplement its services and applications. However, the volume information and network parameters to be optimized grow exponentially, increasing management and orchestration complexity. In this context, solution based on machine learning and artificial intelligence (AI) may be helpful allies in handling this challenge [36, 37].
Global Scalability. One of the most significant advantages of cellular IoT is global availability, allowing fast penetration of massive IoT applications in many markets. However, competing unlicensed technologies can leverage extensive coverage. In special, non-terrestrial networks, such as low-Earth orbit satellites, unmanned aerial vehicle (UAV), high-altitude platforms, have emerged in recent years as strong competitors, particularly for rural, off-shore, and overall remote connectivity [38, 39]. The combination of mMTC and non-terrestrial networks seems to be a vital component of the next generation of cellular systems, aligning with the idea of holistic massive IoT, as discussed in [35]. Some recent works have already shown promising results in this direction; for instance, [40] analyses the performance of mMTC and satellites in a smart city context, while authors in [41] propose a direct connection between terrestrial MTDs and the satellite. Extending this idea, [38] conceptualizes two new architectures to integrate mMTC and satellites, with a direct and an indirect connection between these two entities. These promising findings evince the feasibility of these options even with typical current LPWAN configurations.
1.4.2 Sustainable Connectivity
This section identifies technologies that are the building blocks of sustainable information and communications technology (ICT).
In 2015, the United Nations (UN) defined 17 intertwined sustainable development goals (SDG), while calling on society to meet them by 2030. The ICT and wireless networks industry is one of the most prominent industries to positively contribute towards all established goals, facilitating the creation and access to new services and reduced CO2 emissions [42]. In this context, sustainable connectivity is fundamental toward increasing efficiency in using resources in future networks, as pointed out by a group of international experts led by the 6G Flagship framework [43]. The UN indicates in [44] that the ICT sector is instrumental in making a significant impact on emissions reduction.
Recently, the European Commission announced an overarching objective towards a green future that is economic growth decoupled from resource use [45, 46]. Furthermore, the European Commission has prioritized digitization [45] and achieving carbon neutrality through the Green Deal [46] as crucial development goals for a sustainable Europe. The European Union is setting ambitious targets for climate change and sustainability for the coming decades, accompanied by many other countries worldwide, as evinced by the 2020 Climate Change Performance Index Results report [47].
In line with this, in 2021, the Finnish Ministry of Transport and Communications announced its climate strategy for the ICT sector [48], which comprises six objectives and measures paving the path to achieve ecologically sustainable digitization. Interestingly, some of those objectives resonate with the ideas in this book, namely: 1. increase of the efficiency of resource use, 2. boost of the devices lifetime, and 3. reduction of energy consumption by the ICT sector.
The ICT sector consumes from 4% to 10% of global electric power, and it is responsible for 3% to 5% of greenhouse gas emissions [48]. Moreover, a recent study indicates that devices and infrastructure are responsible for about 70% of the total energy consumption in the ICT sector [49].
In light of the SDG and climate action, many stakeholders in the ICT sector have taken action. For instance, two of the largest manufacturers, namely Nokia [50, 51] and Ericsson [52], have set clear targets towards sustainable ICT. Moreover, along with the UN’ forecast [44] concerning the use of ICT solutions, Ericsson’s research [52] recently predicted a 15% reduction on global carbon emissions by 2030 and decrease to less than 2% on ICT’s global carbon footprint.
All in all, ICT plays a pivotal role in sustainability and climate action. However, the ICT section is undergoing a sustainable transformation as well, evinced by energy consumption reduction, energy efficient designs, inclusion of renewable sources [44, 51, 52]. However, devices and infrastructure are still critical contributors to the energy footprint of the ICT sector. Therefore, it is essential to identify technologies and methods that provide energy-saving solutions lowering their footprint. In this line, we identity key technology enablers toward sustainable ICT. We briefly summarize some of these fundamental technologies, and then in Chapter 2, we delve into technical details and analysis.
Intelligent Reflective Surface (IRS). The IRS emerges as a promising alternative solution to reconfigure the wireless propagation environment via software-controlled reflective meta-materials. The IRS builds on numerous low-cost, passive, reflective meta-material elements. These elements induce controlled phase and amplitude variations on an incident signal. Hence the IRS enables a programmable and controllable wireless environment. This controllability empowers communication engineers to better combat the impairments of the wireless channel in a completely new form [53-55]. Moreover, the IRS does not require a complete transceiver chain. Besides, assembly is flexible and can cover arbitrarily shaped surfaces. These two characteristics impact directly cost and energy consumption.
The transition towards high-frequency bands, namely millimeter and sub-millimeter-wave bands, is underway in cellular systems. In this context, the IRS IRS is an option to provide low-cost link diversity or strengthen line-of-sight (LOS) channel components.
Nonetheless, to achieve scalable performance and overcome competing technologies, the number of elements in an IRS needs to be large. This approach is unbearable in a fully programmable system, since each element requires an individual (phase and amplitude) control, and channel measurement, thus imposing communication overheads. AI-based solutions can help mitigating such challenges [54]. Another option is based on channel state information (CSI)-free solutions (discussed in Chapter 2) to reduce the signaling and instantaneous estimation of controllable features.
Tiny Machine Learning. Tiny machine learning, known as TinyML, is a flourishing field devoted to extremely low power, always-on, battery operated devices, which is stereotypical of many MTDs. TinyML objective is to integrate machine learning mechanisms into microcontroller-equipped objects. The idea comes from the fact that as the network and technology evolve, information processing and AI become more distributed. Therefore, it moves from the cloud to the edge, and now from edge to devices. Notably, it goes toward extremely low power devices, which can potentially make AI ubiquitous for many massive IoT use-cases. Besides, TinyML