Frank Ploumen, is the chief technology officer for IoT platforms and applications at Nokia. Here, he tells George Malim, that among the many challenges facing the further development of the Internet of Things (IoT), technology is no longer the greatest. Technology and connectivity choices in particular still play a critical role – and important work remains to be done – but it is taking a more comprehensive approach to enabling profitable business models and achieving operational excellence that is now at the top of organisations’ minds
George Malim: What do you see as the key challenges associated with opening up new IoT revenue streams for customers?
Frank Ploumen: The biggest challenge with new revenue streams is that people haven’t sorted out what data they can monetise and what data has privacy sensitivities around it. The market is a Wild West environment. Out here in the Silicon Valley where I live, companies have made many trials and there have been many failures. That experimentation is good but monetisation of data in general remains undiscovered territory.
GM: Connectivity is a vital part of IoT. How are the many new connectivity options helping organisations generate profitable revenues from IoT?
FP: Connectivity is mostly radio connections in IoT, there’s very little wired connection. The basics that need to get sorted out are to identify what is the most economic way of collecting the data. From our experience, it’s clear that there is no one-size-fits-all solution and I fully expect many radio technologies to co-exist for many years.
That’s because of the difference in applications and the deployment terrain. Demands change for global deployments compared to a single city roll-out of an IoT application. Similarly, a smart city project has very different demands to deploying IoT in an oil and gas or mining environment.
I don’t think we’ll see a single technology coming to the fore quickly because of this disparity. Having said that, there is a theme with the low power wide area network (LPWAN) technologies that are emerging. These networks, along with 3GPP technologies such as narrowband IoT (NBIoT) and extended coverage GSM, focus on extending range and maximising battery life. The only variable is bandwidth availability and spectral efficiency.
GM: Does the somewhat bewildering array of technologies hinder rather than help?
FP: In the short term it’s a bit of hindrance in the sense that there’s a lot of confusion and misinterpretation in the market. However, I look at these different technologies as items in a toolbox from which organisations can choose the most appropriate system for their deployment, balancing capex and performance. Non-experts, though, will find this choice less clear so I could see five different technologies compete for the same value proposition.
The other issue to consider is the lifespan of the technology selected. If a city deploys sensors for monitoring traffic, parking or the environment, the chances are it is looking for the deployment to have a life of 10, 15 or 20 years. Given that, it’s probably not realistic to be able to make a safe decision on a technology to be out in the field for that length of time. Therefore organisations need to create an environment in which it’s economically viable to change technology. Connecting a traffic sign to cloud today can be very different to how it was ten years ago so organisations need a way to mix and match technologies so they don’t lose their entire investment.
For radio connections, the main challenge is that people complain about their cost for IoT. If we can get the customer to a price point where a 3GB package goes from $20 down to 3MB for two cents the cost is no longer a dominant issue. The challenge is that we can’t do that with today’s radios but, by losing track of cost, we’re forgetting the benefits of existing radio offerings in terms of their security, ultra-reliability and their global footprint. A lot of LPWA discussion centres on reducing the cost but in doing so we must keep an eye on what other attributes could be lost.
GM: To what extent is the sunsetting of 2G mobile networks an opportunity to move to a more appropriate connection technology for each app and service? Is there a silver lining to 2G retirement?
FP: The sunsetting of 2G is the first real world scenario in which we need to deal with the fact that radio technologies don’t stay around forever. Replacing radio modules is probably not the key heartburn – although the labour cost may be significant. The real cost is for apps that were designed for 2G and the integration of systems that were designed to be set and forget.
This is a significant challenge but the industry is learning from it. I don’t believe that you will find an operator that wants to give a 20 year guarantee on network availability but, if you want hardware for long term deployments such as highway signs, you will need to be prepared to change technology at some point. The only way to avoid such a change would be to build a private network but the economics of building one don’t stack up for that either.
An important point is that 2G was never designed for IoT, it simply happened to get repurposed for IoT. New technologies like NB-IoT are purpose-built or adapted to be fit for purpose for IoT so I expect them to have a longer life. By relying more on IoToptimised radios we’ll have less dependency on a sunset dictated by external factors in future.
There’s also the option of taking a hybrid approach such as building a solution based on new technologies and, when that sunsets, there’s nothing preventing you from converting your system to a private network. The choice then becomes whether to take on the burden of keeping the radios alive in your private network or replacing all the sensor modules. You could select technology today and plan to turn it into a private network in ten years, it becomes an economic choice at that point in time.
GM: What is Nokia’s view of the low power wireless solutions that are coming to market? There are clear use cases but do you see them causing fragmentation and resulting in lack of viable standardisation?
FP: There’s definitely some fragmentation but over time we’ll see a set of radio modules emerge that are multi-standard. For a lot of radio technologies, when you look at the physics, there’s no reason why they can be multi-protocol or multi-technology based. It just becomes a software question at that point. Obviously for IoT to be a success, fragmentation needs to be addressed. We don’t want 12 different radio modules available.
GM: What do you see as the most important ingredients organisations need to enable them to get to providing new IoT services quickly?
FP: When we look at IoT applications the work we do is two-fold. We build infrastructure and platforms and targeted applications where Nokia is active. The temptation is to build something lean and vertical but, while that’s really fast to prototype, you quickly run into problems. The ugliness of going from the prototype to real deployment lies in other boring and mundane aspects such as the authentication. The best way to accelerate development and have a fast market introduction is to use known expertise and that drives you to a horizontal platform approach.
If you take the example of a TV set top box, you’d pick a Linux operating system and build on top of it. You’re not going to develop a new operating system even though you could. You would save integration time by doing so but you’d soon get into trouble because you’d need to create a filing system and administration privileges, among others. The trade-off lies in using proven components at scale which I think is the biggest opportunity to accelerate IoT deployments.
GM: What do you see as the main bottlenecks that are slowing down service deployment?
FP: IoT is very much a partnership or community environment. Very few organisations can build solutions from A to Z on their own and, if they do, be very afraid because they’re probably locking you into proprietary technology for life.
Bringing all of these components together doesn’t happen by itself. Systems integrators tend to make money by bringing components together but, as the industry evolves and more organic partnerships happen, things should work out of the box. If an organisation wants a best of breed solution it is not unrealistic to ask for Watson on top of a Nokia platform, for example. This type of outof-the-box availability will happen and bring costs right down.
GM: How can Nokia help customers address these?
FP: The platform we have built is driven by open standards and modular interfaces. We’re trying to push a modular architecture where we have a clean separation of functional blocks. Some people see this as overly complex because the architecture has a lot of blocks and interfaces but we recognise customers may have already invested in some of these and we don’t want to push a onesize-fits-all agenda because then we are back to selling point solutions. We’re working to ensure this is an open ecosystem with welldefined building blocks.
In 2016 we set up The IoT Community to foster the open, collaborative environment that IoT requires. We now have approximately 300 companies active composed of hardware and software vendors and municipalities and telecoms operators. There’s a very wide spread of companies involved and they’re focusing on use case development and how to make the business cases work.
The technology isn’t the highest risk component, the biggest unknown in IoT is how to make money with use cases and that’s much more difficult to predict than how technology development will play out.
GM: Given that the platform or foundational technologies of IoT are widely and readily available, how can organisations differentiate their offerings?
FP: A lot of the underlying technology is readily available and open source so the biggest opportunity to differentiate comes down to operational excellence. It’s not hard to connect a streetlight to the cloud but what we can do to differentiate that is to provide excellence in scalability, reliability and security.
The applications that sit on top of infrastructure ultimately are differentiated by their creative use of data so the quicker we can get to the separation of responsibilities the better. A data scientist wants to have as little as possible to do with the infrastructure. They just want the data to be delivered reliably and so they can analyse it and create differentiation.
There are several standard elements that must be addressed to achieve operational excellence. These include creating the right cost structure, making the right choice between private, public or multi-tenant cloud and deciding whether to use edge computing or not. These are all choices that must be made within the context of specific app or service and from an infrastructure point of view we need to give companies the right tools to give them these clear choices. Only then will they be able to have their data delivered at the right time, at the right cost and with the right service level agreement.
If I was an application developer, I’d say the compelling business case for an IoT infrastructure is guaranteed security that is run at the cost structure required by the app. If you can do that, you can enable someone to differentiate their offerings.
IN ASSOCIATION WITH NOKIA