Tech-Talks Bregenz | Apr 14, 2016

Tech-Talks BREGENZ - Karl Jónsson, IoT Architect, Tridonic

Karl Jónsson, IoT Architect at Tridonic presented some interesting facts, figures, and ideas about IoT and the future of lighting in his speeches about “The Internet of Things and Lighting” at Zumtobel’s International Year of Light Event in 2015 and again at the Light + Building 2016 preview. LED professional got the chance to talk with him about his visions for IoT and Tridonic’s latest developments in that domain.

LED professional: Welcome to Bregenz, Mr. Jónsson. As a short introduction, could you please tell us what makes IoT so interesting and why lighting will become such an important part of it?

Karl Jónsson: Yes, of course. IoT offers many new opportunities to established infrastructures, especially for lighting to become the connectivity backbone for IoT sensors and devices to offer new services and automations. For example, in a big office building there might be a few hundred or even a few thousand bathrooms that have to be checked every day. Soap dispensers and napkin holders need to be checked and refilled. They could have built-in sensors to connect to the light and signal that they need to be refilled. That is just one example of what can be monitored. Retail applications is another example where digital price tags need to be managed or customer behaviors monitored in order to offer improved services. Space management and remote monitoring of energy is also a use case that is high in demand where IoT integration with lighting will be a great benefit.

IoT needs a dense network to allow these services. While we mainly use our cell phones as a bridge to the world for our apps in our private and consumer life, this is not as easy in the professional environment. Therefore, another bridge needs to be used. Light is the densest network of wired devices worldwide and this is a big chance for lighting to become the infrastructure and a core enabler for this technology.

LED professional: That means that this would go beyond lighting tasks. It would become a data collection resource, and an infrastructure for other devices. What are the main requirements to allow these new opportunities and to get it to work smoothly?

Karl Jónsson: The first thing of importance is standardization. It has been a struggle in the past and it is ongoing. But it seems that the industry is looking more and more towards the internet protocol (IP) and moving away from individual solutions.

Screen Shot 2016-04-14 at 09.13.42.png

LED professional: This requires new developments which draw huge R&D resources. One could argue that any lighting task, sufficient control and monitoring of the lighting infrastructure could be fulfilled by connecting DALI to the web using a gateway. How would you counter that argument and what were the reasons for joining forces for this new development?

Karl Jónsson: DALI and other non-IP-based protocols are certainly efficient and good enough for today’s use cases. The upcoming problem, however, is that they operate in silos and are self-contained without the possibility of interconnecting with pier building automation systems without translation. This is the case especially when it comes to support sensors, data collection and other features through the lighting network that are beyond today’s use cases.

This is also true for many other systems. Lighting, building automation, security systems, retail services and other professional services still operate in vertical silos and in independent ecosystems. The operation of these systems is, therefore, very inefficient and the systems are not scalable. Creating a joint ecosystem and making systems “speak” with each other “over the top” through cloud based wrappers and gateways can go a long way, but doesn’t guarantee security, reliability, operation efficiency, and, in the end, customer satisfaction

LED professional: Can you go into detail concerning silos?

Karl Jónsson: The best analogy I can give is the cell phone market. Two years before the first iPhone came out, I visited Nokia innovation lab in Oulu Finland where I witnessed a mind-blowing application and development around the Nokia Symbian OS. As an example, they demonstrated an application that used the phone camera to translate a Chinese menu to English on the fly, so the display showed the English text. There were also payment solutions and home automation use cases that, at the time, felt unreal but are a commodity today. I believe the reason why this never became public domain at the time was not only because of the aesthetic design, user interface and business model as often referred to. I think it is much more because of the same technology silos we are trying to avoid today, which are the closed platform, a limited update strategy as the entire phone needed software update when new “apps” were added, and the infrastructure dependency.

Screen Shot 2016-04-14 at 09.16.21.png

Closed or open can mean many things. For example, you could argue that Apple is a closed ecosystem, but it’s open enough that protocols are a commodity and providers can compete on the application layer. It will be hard to convince an HVAC manufacturer, for example, to support DALI. Therefore, there is an increasing demand for a common networking layer for M2M communication and centralized operations. If everyone communicates over IP, it’s easy to enable interactivity both ways, when the business calls for it.

LED professional: Could you give us an example of why it is such a big advantage to use the IP-based technology approach?

Karl Jónsson: Adding a new feature in today’s systems can be very difficult. For example, let’s say we wanted to add a microphone that detects sound patterns for context awareness in a room like a meeting, talking, vacuum cleaning, a party, and so on. We would have to get DALI alliance to support such features or try to be creative with existing standards with a proprietary modification. Then we would need to push a firmware update to the gateway in order to support the feature. At the gateway the connection is terminated for protocol translation, which could give hackers direct access and let them eavesdrop on the audio coming from the microphone. The communication protocol between the gateway and cloud, which is commonly proprietary, would have to update and finally the cloud, apps and application interface would have to add support for this. Not to mention QA that would have an exponentially growing test matrix of features to guarantee stability instead of just having to test the device and application itself.

In an IP world, the device would be added to the network and be told which infrastructure device, in the form or a switch or access point, it is allowed to talk to. From there, only the application and business logic needs to be updated at the other end as IP-2-IP security and interoperability is already a commodity.

LED professional: Does that mean that the IP protocol is the way to go for IoT and smart lighting for Tridonic and the Zumtobel Group?

Karl Jónsson: Yes, for the whole group it is all about following standards. And the Internet protocol is therefore very crucial for us. Currently, there are technologies offered that are bridged into the IP world that have their own benefits, like DC-String and ready2mains which are part of the company’s ecosystem. But the most straightforward approach for now is net4more.

LED professional: Tridonic also belongs to the OpenAIS consortium. How is net4more related to the OpenAIS project? Is it a direct outcome of this project?

Karl Jónsson: With the development of net4more, we already started R&D activities and plans to move into IP-based lighting. Therefore, we were very happy when the OpenAIS project was initiated. Since we believe that this IP technology will pick up fastest and to the best customer satisfaction, we followed the idea to design it right from the start as open as possible. We were happy to see that the experts in OpenAIS, coming from lighting, building automation, semiconductor and IT, came to the same conclusion on how a system should be designed. And some new ideas where born!

While net4more is not a direct derivative of OpenAIS, it shares synergetic properties that we are trying to drive and align between other standards as well. We see net4more and OpenAIS moving in the same direction but at different speeds and with slightly different scopes.

In OpenAIS we hope to fill all the gaps that we have already identified in the available technical standards and technologies. Professional lighting has different requirements than the telecom industry. Lighting needs peer-to-peer communication and multicast messaging. It is also different from the home automation and, therefore, the current Thread Group specification 1.0 does not fulfill all lighting needs, like commissioning of big installations with hundreds or thousands of luminaires. In OpenAIS we do research on several layers of the communication stack in the field of security to solve these issues.

LED professional: Does that mean that customers have to wait until OpenAIS has delivered in order to benefit from an IP-based lighting solution, or is net4more already future-proof?

Karl Jónsson: No, not at all. We designed net4more as a starting point of a future-proven and upgradable smart lighting solution that is based on open standards where they exist. Once the OpenAIS solution becomes a standard, we will be able to change over and merge both solutions easily.

We are an active member of several consortia and committees (IETF, Thread Group, Fairhair, TCLA) where we want to contribute to a fast standardization process for IP-based lighting together with players from the lighting, IT, and building industry.

LED professional: Could you please give us an idea about the cornerstones of this IP-based technology and how it compares to other approaches?

Karl Jónsson: It is important to be based on end-2-end IPv6 technology, which is IPv6 for low power networks, and in case of wireless 6LoWPAN. Other cornerstones are that it offers a common, secure and open way to connect on-board and pair and secure wireless devices to a network. It is also important that no gateways are needed for native net4more devices, which offers many benefits compared to conventional solutions. For instance, scalability is not limited to a proprietary addressing scheme as IPv6 has practically “unlimited” addresses available. It is based on open standards where available. No protocol translations are necessary, and hence, infrastructure devices do not need updates to support new features. The system relies on common IETF-based standards that are already empowering today’s Internet of over 8B devices, and when needed, it can virtualize DALI and other non-IP-based devices into IP.

LED professional: Not all of our readers might be familiar with the IPv6 specifications and related network technologies. Could you explain the difference between IPv4 and IPv6 and tell us some of the benefits?

Karl Jónsson: IPv6 is the 6th version of the popular Internet Protocol. Although IPv6 is becoming more dominant, IPv4 currently powers most of the 8B devices in the world today, like computers, phones, and tablets. The problem with IPv4 is mainly its addressing scheme, which is limited to 4.3B addresses. Because of this, many networks, especially home networks, are currently using Network Address Translation (NAT) which allows all devices in a home to connect by using a single IPv4 Address. IPv6 is backwards compatible and solves the addressing problem by allowing 3.4x1038 addresses, which should never run out. Devices can thus avoid any translation and talk directly with each other when allowed and authorized.

LED professional: Both net4more and OpenAIS support wired or wireless communication. Are there different target applications and safety issues?

Karl Jónsson: The architecture of an IP-based lighting system should always be designed to work independently of the chosen physical layer. This is a big progress compared to DALI where the protocol is entangled with the method to transmit it. Therefore yes, both wired and wireless connectivity methods are supported. While both will be very safe and secure, they will have different benefits depending on the installation and application. As an example, in a public place, it might be more appropriate to use wireless solutions to avoid people unplugging an ethernet cable to gain access to a local network. In a hospital, a wired solution might be more suitable to reduce wireless noise for potentially sensitive medical equipment.

LED professional: PoE is currently another trend and a relatively new approach for Tridonic. We’ve learned that currently, approximately 60 W can be supplied and a new spec to allow up to 95 W is in progress. Is PoE also an integral part of the net4more concept? Is this specification really sufficient for all applications?

Karl Jónsson: Yes, PoE is also being considered. It’s ideal for IT-based installations where lighting needs to be a part of the IT network. But it is not ideal for high power applications like high bays, stadium lighting or most other outdoor lighting applications.

LED professional: This means there are some specific target applications. What are the greatest advantages?

Karl Jónsson: It’s mainly office panels and downlights that are being targeted. For these applications, PoE offers a strong value proposition when the lighting control system is designed like an IT system. IT planners are very familiar with this technology and it allows cost savings during installation.

LED professional: One is still fighting for any percent of increased efficiency today. But PoE has relatively high line losses because it originally was not intended for transporting that amount of energy continuously. Isn’t that a significant disadvantage?

Karl Jónsson: There are pros and cons. Some or all of the line voltage can be regained by decentralizing the PoE switches that shorten the cable that runs between luminaires and switches. This lowers the line losses. Intelligent light scheduling of such intelligent systems will also contribute to savings.

Screen Shot 2016-04-14 at 09.18.58.png

LED professional: All these discussed systems are IP-based. People often associate security and privacy issues with the Internet and IP. Adding these sensors and actors from their currently “private” environment directly to the Internet could cause some uneasy feelings when they think about all this information being in the cloud. Can you tell us about security and privacy?

Karl Jónsson: Looking at the pitfalls and history of smart lighting, we see that until now usability is one of the most relevant topics. But it has become much less an issue with the introduction of the smart phone, which we all know how to use. While today it is protocol and interoperability, tomorrow it will be security, data collection and privacy. This cannot be an afterthought. We have to think about it now.

Looking back at the history of data breaches we see that a lot of common mistakes can be avoided. When you are running on the IP protocol and you follow the common standards and practices in security, you are pretty safe. Please note that I am talking about security and not privacy or data collection. If you get audited by one of the major security companies it doesn’t necessarily guarantee that you won’t be hacked. You also need to look at the history of recent attacks. These attacks are not usually very innovative in that the same tools are usually used. Looking at the latest trends at hacker conferences like the Black Hat in Las Vegas might help to minimize risk.

LED professional: In a speech you pointed out that the biggest safety risk is the app, not the system. How could this risk be minimized?

Karl Jónsson: In this regard, the biggest risk is most often the user of the solution. Very often they don’t change default user names and passwords like admin/admin, or they use simple passwords to easily remember them. Furthermore, the end device, either PC or smartphone, is typically an uncontrolled domain adding the risk of spoofing and backdoor software compromising access.

There are several ways to mitigate the risks. One of the best solutions would be to get rid of passwords and use a secure token from an NFC care, phone or fingerprint. On another level it would be helpful to design the system in a tiered approach so that one compromise only affects small parts of the systems. That could mean to build the security architecture as a “marketplace” instead of a “castle”. And last but not least, simply following known standards on Internet security and to comply with the security and privacy regulations would mean a significant reduction of risk in most cases.

Since bank transactions, aviation controls and other critical applications can rely on the IP infrastructure, lighting can, too.

LED professional: This all is about security, but what about privacy?

Karl Jónsson: Privacy issues depend strongly on where a solution is located. It is, in many aspects, a legal issue. Germany has probably the most stringent law. Many German companies don’t allow the use of US-based hosting services. They are worried about data collection. But there are ways to keep it anonymous. One popular trend is fog computing where data are preprocessed on site. You are not sending raw data but just relevant data, already stripped of the privacy part. For instance, from a camera for people counting, it only sends the numbers and not the actual pictures. Such technologies will become very important. One relevant side benefit of this technology is that the amount of transmitted data is drastically reduced.

There is still a certain disconnect between what is anonymous and what not. For example, when you are walking through a supermarket, and you are tracked based on the ID of your phone, all actions and reactions are collected, even when they are not tied directly to your personal information. Is that anonymous or not? You are still targeted as your personal device.

LED professional: Could we conclude that this means that the exact meaning of privacy must be agreed on and is necessary to develop or chose an appropriate technology?

Karl Jónsson: Yes, that would be the ideal situation.

LED professional: Earlier in the interview you mentioned two systems that Tridonic developed and introduced during the last few years: ready2mains and DC-Grid, which are bridging the IoT world. Are they equally well suitable for IoT and when are they the right choice?

Karl Jónsson: There are applications with less sophisticated requirements where costs are more important. Both are also developed for professional applications, but their IoT capabilities are limited. They can bridge into DALI, or later, net4more, but end points cannot feed back any information and they cannot control lights individually. Therefore, integrated sensors are not supported. They offer limited applications aside from dimming and energy savings.

LED professional: Which technologies are used for these product lines?

Karl Jónsson: These systems use the power cables for communication but have distinct technical differences to Power Line Systems and should therefore not be confused with them. For the DC-String no common standards were available because there are just a few activities in the DC domain, and most of the AC power line activities are related to high data transmission rates that are not necessary for this application. For us, stability is much more important. Therefore we had to develop our own technologies and communication standard.

LED professional: What were the major reasons and applications for developing these products?

Karl Jónsson: In the case of the DC-String it currently benefits applications where space is limited for luminaire designs and individual controls are still desired. With DC-String the power supply and control can be centrally located without compromising these requirements.

Moving forward, we see a potential trend that buildings will be designed completely off the grid and become CO2 neutral. Tesla’s announcement of the power wall and power pack for residential and industrial buildings could accelerate this trend. DC-String has big advantages for such buildings. It does not make sense to convert DC voltage from solar panels to AC, then again back to DC for batteries and then again to AC for the lights because LEDs need DC.

Ready2mains uses a very lean way to send digital commands on the mains' wires to enable group switching and dimming via a gateway. For retrofit from conventional to LED light where no granular light control or management solution is required, ready2mains would still be the best choice if cost and energy savings are the primary goal.

Screen Shot 2016-04-14 at 09.21.40.png

LED professional: Besides end-user functionality and costs, there is another important aspect: Installation and commissioning. I know that one of Tridonics’ goals is to develop a solution that also works for “uneducated” users. This raises several questions. What would this type of system or approach look like and how would it work? Is it possible to keep it simple while being secure at the same time?

Karl Jónsson: A detailed explanation would probably break the mold, but in general the millennial generation of installers and commissioners will have limited tolerance for a learning curve and will expect things are WYSIWYG. With this upcoming new generation of professionals in mind we are exploring a new visual and fun way to perform commissioning without compromising on security and still allowing experienced professionals to familiarize themselves with new and exciting features to improve and speed up the commissioning process.

LED professional: As you mentioned at the beginning, one of the most exciting issues of IoT is that it is at the starting point and may open unknown future prospects. At the same time, there are other technologies emerging like Li-Fi. If you don’t mind, could we concentrate on these prospects and visions in the last part of our interview?

Karl Jónsson: Yes, of course.

LED professional: Could you please give us an idea and update of the future of IoT, related technologies and their applications?

Karl Jónsson: Aside from remote monitoring and space management, there are several other services that could benefit from the net4more infrastructure, only the imagination is the limit to what applications could be built once the lighting network is IP-enabled. Another example of driving service enablement is certainly Li-Fi where light itself becomes a physical layer in the networking stack, which would be a natural extension of an IP-based luminaire when each luminaire becomes a Li-Fi access point. Li-Fi is focused on high-speed data transfers and is targeted to compete and compliment WiFi. The current problem with Li-Fi is that it requires a dedicated receiver on the user side so it won’t be popular unless a popular device manufacturer decides to natively integrate it in their devices. However, there are already strong rumors this might become natively supported in future iPhones. That could be the breakthrough for this technology.

There are, however, other lighting modulation schemes, that operate at a lower frequency so a smart phone camera is enough to receive the signal. Generic light modulation is typically referred to as Visual Light Communication (VLC). While low speed VLC is not useful for high-speed data transfers, it is good enough for other applications such as indoor navigation and for localization during commissioning. VLC is typically used in conjunction with a smart phone’s Gyro and magnetometer to determine place, angle and orientation.

LED professional: You once said, “The true benefits of IoT may not be that much for already known applications, but rather for new ones that are currently not yet known.” Could you give us examples?

Karl Jónsson: IP-based lighting is often referred to as the “backbone of the IoT”. A meshed wireless lighting network could give other “things” access to the Internet. Things like distributed access points that allow small-constrained battery-powered sensors or actuators to communicate with the Internet amongst themselves or with the cloud, even if their information is not needed for lighting. It could be used for logistics, climate control, safety or asset tracking. This opens up new service possibilities that can be hosted by the platform created through a connected lighting network.

Other examples I can think of are the integration of the lighting backbone with other emerging technologies like Augmented Reality and self-driving cars, where lighting could play a role in context awareness and communication.

LED professional: To close, could you tell us what is the most exciting future prospect for you, personally?

Karl Jónsson: I’ve been obsessed with lighting ever since I can remember and my personal passion is when a light source can be taken under complete control and the power of darkness can be appreciated. I’m excited about making lighting truly dynamic, something that’s not just a statically mounted device that offers services, but a truly integrated experience that can become a part of my digital life.

LED professional: Thank you very much for your visions and insights on IoT and Tridonic’s latest activities in that domain. It was a pleasure talking with you, here in Bregenz?

Karl Jónsson: My pleasure.

Karl Jónsson
Karl Jónsson joined Tridonic in 2014 as IoT Architect to pave Tridonic’s way into the Internet of Things era. Over the last 6 years, as the General Manager for the Smart Lighting business unit at GreenWave Systems^ which he co-founded in 2008, he pioneered the world’s first IP based lighting solution for consumers. Before that, he reported to the CTO office of Cisco Systems consumer group. In the early 90’s, he started his carreer in entertainment lighting with over a 10-year track record at major TV, theater and show productions in Scandinavia. Karl Jónsson holds an M.Sc. degree in Information Technology from IT University of Copenhagen, a B.Sc. degree in Electronics Engineering from Copenhagen Engineering College and an Electrical Science Degree from the Technical College of Reykjavik. Karl has over 50 patents filed and granted worldwide in areas of IoT and lighting, and has been an active member and strong influencer of major standards bodies.