The Internet of Things Boxing Match: Cellular versus Non-Cellular LPWAN

Introduction

A vision of a globally connected world, where anything would be able to communicate with everything, while sounds rather silly to some or alarming to others, is in fact one of the mostly discussed topic nowadays in the industry. Estimates of the IoT market for the upcoming few years have been laying around in the industry for quite a few years. The estimated number of connected devices have been reported to range from 12 billion to an astonishing 50+ billion devices. Debates started to arise when analysts and market researchers started forecasting the penetration of different wireless access technologies used to connect these devices. These estimates would definitely impact future investment decisions, and indeed have an influence on regulatory decisions and policy makers.

Whether these reported figures are correct or overly estimated, in my views, that really doesn’t matter. After all, most of these forecasts depend on the way you account for your connections. To give you an example, let us have a closer look at a smart connected vehicle use case. Would you consider the in-vehicle connected peripherals, e.g. temperature sensors, tire pressure sensors, air bags sensors, A/C control, lighting, entertainment...etc., all as the objects or things in a typical IoT system? Or woild you consider the vehicle as whole, a single device or a thing connected to other things (e.g. traffic lights, other vehicles, passengers, pedestrians,...etc.) in a larger system?

Connecting all these things inside the vehicle together can by done by a variety of technologies (both wired and wireless). Connecting the vehicle as a whole with external things can be done using different wireless technologies as well (e.g cellular versus non-cellular). In principle, the IoT sphere embraces wired, and wireless technologies, and the concept spans multiple domains including proximity protocols, access and network technologies, cloud, platforms and services. IoT can be modeled as a complex system (of systems) with different levels of connectivities, between themselves and also to the Internet.

When we analyze the market size of IoT, we usually bump into some key questions, especially when try to estimate the future penetration of wireless IoT access technologies like Low Power Wide Area Network (LPWAN) and cellular IoT. A sample of these questions is on how you can spot a market opportunity if you are running an IoT related business? How do you know which factor(s) is actually driving the market?.. is it the price?.. the price of what? Sensors, or gateways, or other network/cloud components? And what is price sensitivities of these products? Who actually did any price sensitivity analysis for any IoT related vertical market? Would the market supply trigger the demand or is it actually the other way around?

In a way, these questions took me back to 2009 (or even earlier), when the industry and the different administrations all over the world were occupied in promoting and developing their own broadband strategies. There was a heated debate back then on which of these two factors actually tiggers the other, the supply or the demand. On the other hand, impact studies were constantly reported and publicized. Econometric studies published by esteemed agencies like the World Bank, provided estimates for the impacts on the economy on a macro level. Most notably, these studies illustrated that every 10 percent increase in the broadband penetration led to a one percent increase in national GDPs of countries.

These studies were definitely correct, and till this moment, they are considered as rule of thumb for assessing the impact and importance of ICT infrastructure on the development and growth of the economy. But these impact study results don’t come unconditionally. Policy makers and industry leaders, ironically, forgot to highlight some of the very important facts, disclaimers, and assumptions related to these studies. For example, most econometric studies conducted to assess the impact of broadband on the economy, were correlational studies. Economic experts know quite well that correlation does not imply causation. You don’t build a broadband infrastructure and expect all good things to happen by itself. Studies indicated that a key ingredient for the development and economic growth to happen, is the presence of what we would like to call the “absorptive capacity”.

A key ingredient for the success of broadband plans is the presence of the absorptive capacity

The “absorptive capacity” generally refers to is the ability of enterprises, governments, and nations to turn a broadband connectivity to an actual useful product or service. The point to make here is that value comes from the transformation of the connectivity into something useful for the organization or the enterprise. It’s the process of value creation that matters.

History Repeats Itself

Take a look at the current debates surrounding the IoT industry nowadays. Although they are slightly different than the broadband case, the debates can be seen as to inherently have more or less the same principal differing views on how the IoT can reach its full potential.

Two main views: the first is the supply view which sees that deploying an IoT infrastructure would trigger the user’s appetite and would actually trigger opportunities. Would an IoT system bring tremendous savings and efficiencies to your business operations, just by deploying it?

The second is the demand side view which thinks that the demand on the applications and services is the main driver to reach the full potential of IoT. For example, the analytics demanded by the user, and generated by the system could be what actually triggers informed decisions (hence operational savings and efficiencies). So the question here is which if the two perspectives are right. May be both or neither.

We need to look for the “absorptive capacity” of the IoT. We need to research the actual critical success factors of value creation. Candidate factors include the presence of an effective and efficient business process, in addition to the existence of an ecosystem that would cater for ideation, innovation, and business development. In highly regulated markets, with difficult access of development platforms (reconfigurable hardware platforms, IoT cloud based platforms, devices, sensors,...etc.); developers, SMEs, and start-ups will find it difficult to create and innovate and try new product ideas to solve actual chronic problems.

In a such a market, a monopoly allowed to deploy IoT on a nationwide scale, may not trigger efficiencies and cost savings promised to its client, simply because there would be no competition, no ecosystem to serve for alternate products or services. You would only take what you are offered regardless of its technical merit or financial viability.

Sounds familiar, right? Same old stories with the development of broadband, where aspects related to competition, market entry barriers, incentives, regulations and ecosystem are all crucial for the successful deployment of the system. There is a lesson here. Do not repeat previous mistakes. Focus on the ecosystem at large and drop unnecessary debates. Focus on the important questions: how could we deliver value.. real value, to our organizations, societies, and industries, instead of the limited view of how to best sell our products, and how to do that as quickly as we could possible do. May be it’s better to start thinking of how we could sell trust instead of selling analytics!

Sell trust! Do not sell analytics!

Digital tranformation is also critical for developing a wholistic strategy that would employ digitization technologies to create value. To fully leverage the changes and potential opportunities introduced by digital technologies (including IoT), a well crafted transformation strategy is crucial for reaping the full benefits of an IoT deployment. We cannot separate what we do, say an IoT system deployment, for the bigger picture.

LPWAN (Non-Cellular) versus Cellular

Another heated debate in the industry these days is on which proximity/access technology is the best for the Iot use cases. Is it the non-cellular Low Power Wide Area Network (LPWAN) technologies like LORA, SigFox, and others? or cellular IoT like NB-IoT and LTE-M? In principle, the core of the debate lies on the pros and cons of the licensed spectrum versus the unlicensed spectrum technologies. Cost is another critical aspect. LORA and SigFox advocates highlight the tremendous and significant lower cost per module of non-cellular LPWAN modules, which typically ranges from 2$ - 5$ (as compared to the NB-IoT modules which range from 7$ - 10$ per module). Furthermore, the certification costs, is also significantly higher in the case of cellular IoT modules as compared to the non-cellular case (around six-fold higher). So selecting the optimum connectivity technology depends on many factors, most importantly the use case requirements and costs. Requirements alone are not enough!

But connectivity shouldn’t be our biggest concern, and to illustrate why, it’s worth noting that connectivity only accounts for about 4% - 5% of the total value chain of a typical end-to-end IoT service. That’s why, debating which technology is best; is absolutely non-sense in my views. We shouldn’t also forget that interoperability and interworking is one key aspect for the successful and fruitful implementation of an IoT system (i.e. to achieve its intended results).

Generally speaking, we keep hearing long discussions and arguments on which technology is better for the access, and which is better for the transport. We are forgetting a very important thing. When we talk about IoT, we usually coin it with automation, intelligence, analytics...etc. It’s as if we want to say that a system thinking on its own, acting on its own (correctly) is the ultimate thing. We are blessed with the gifts of rationality, volition, and feelings. We can acquire taste, and lay our own judgement. We must think of how we can code these aspects into machines, and how to unify our understanding of the best business processes for a specific application or use case.. and here is where there is an imminent need for standards. Not only does standardization play a critical role in developing interoperable solutions, unifying requirements, architectures, and specifications; but also helps in reaching a common understanding of the processes needed to manage a specific domain or vertical.

Among other notable standards developing organizations, the ITU-T Study Group 20 (SG20) plays a role in filling the gap in some of the above aspects. SG20 is working to address the standardization requirements of IoT technologies, with an initial focus on IoT applications in smart cities and communities.

SG20 develops international standards to enable the coordinated development of IoT technologies, including machine-to-machine communications and ubiquitous sensor networks. A central part of this study is the standardization of end-to-end architectures for IoT, and mechanisms for the interoperability of IoT applications and datasets employed by various vertically oriented industry sectors. Additionally, it studies the use of ICT infrastructure and relevant models such as implementation and deployment models, to ensure end-to-end connectivity and service management, using a variety of technological domains. Standards are critical for the development of our industry. It’s not a luxury, it never was and it will never be.

Conclusion

Some takeaway messages:

• Digital transformation is crucial for a successful and wholistic IoT deployment strategy. Remember, it not about the technology, it’s about what create most value for the user and how. And that can only be achieved if you took care of all the tiny details including the non technical aspects like optimizing the business process.
• Standards are not a luxury. Standardization efforts are needed at all levels of the value chain to realize promised gains.
• Cellular versus non-cellular IoT access technologies.. there is a room for everyone. It’s the use case, and the cost that would finally assist us in selecting which technology is more suitable for a particular use case. But remember, the whole connectivity part accounts only for around 5% of the whole value chain!

————————————————————————————————————

Ramy A. Fathy is the Vice Chairman of ITU-T SG20 on IoT and its Applications including Smart Cities and Communities. With 15+ experience in the digital services space, Fathy has been involved at the different stages of development of various standardization and digital transformation initiatives. He's an entrepreneur at heart, and a senior lecturer on topics related to IoT and AI.