LyfeNet Solutions Pvt. Ltd is a Chennai based IoT company serving Smart Industry and Retail segment through its Consulting, Product Designing & End–End implementation Services. Mr. Mohan Kumar Sundar, Director at LyfeNet Solutions throws a light on the technology trends emerging in the world of IoT.
Q. In your opinion, which are the top 3 technologies that are shaping the future of IoT? How are they influencing the future of IoT?
The list of key technology drivers is longer, but let me highlight some of the areas that are shaping the future of IOT.
- Connectivity: Connectivity is the backbone of IOT. Optimal range, be it short range or long range network, low power and support for good device density, and bandwidth. This is would be an ideal world ask. Today there are a large number of competing and evolving technologies in fray, like the NB-IOT/LPWA, LORAWAN, Sigfox, 802.11AH, 802.15.4, Bluetooth 5.0 etc. However, there is no one single dominant technology that can fulfil all these asks. Hence commercial and technical trade-offs need to be made. It is not just the technology but also the development the eco-system around it that will determine its adoption. This is going to be an interesting space to watch out for.
- AI/ML: Ultimately all the data that IOT generates needs to make sense, help increase productivity, efficiency, make real-world decisions and make business sense. It would humanly impossible to churn through the huge volumes of data generated and manually derive sense out of it. AI/ML fits the bill. IOT and AI/ML will be like Yin and Yang, creating a whole new world of possibilities.
- Platforms: On the hardware side of platforms, the growing need for platforms that come with processors packing higher processing punch yet low on power consumption, footprint, capable of real-time processing, along with infrastructure components like the communications, interfaces for data acquisition, security, encryption, firmware updates etc. This is also key in driving Edge computing capabilities.
On the software side, development of numerous software platforms providing protocols, middle-ware, horizontal/vertical platforms etc.., analytics and AI platforms, help to quicken the pace of development of IOT solutions. Again there is no single end-to-end platform, but these provide us the important building blocks.
If I may add one more key area, it would be:
- Sensors: The applications of IoT is shaped by what sensors can sense. And in recent times there has been an explosion of sensors development. And are no longer just simple sensing devices but are pretty complex, low powered, tightly integrated miniaturised devices. Take for example personal radar that can see in 3D, costing a few dollars and fit into less than 1 sq inch, opening up a wide range of applications that would not been possible earlier, right from automotive to healthcare.
Q. Between the cellular and NB-IOT/LPWA families of wireless technologies–which one are you betting upon? Why?
There will not be one single technology that will dominate or will become de-facto technology. In fact, we have to look at it as 2 parts, 1. Local area connectivity, 2. Wide Area connectivity. Of course, there will be grey areas where Local Area connectivity can play the role of Wide area connectivity and vice-versa, as in the case of Wi-Fi – 802.11AH, which can provide connectivity to even a kilometer range.
But by and large the choice of connectivity will depend on the use case, say for example, if the use-case is to connect PLCs’ of 100s of machines in factory churning data at sub-millisecond rate, then arguably WiFi is a better option, simply because the need here is a local network rather than a wide area network, that is capable of supporting high-speed data rates and high device density. However, if the use case is an environment sensor that requires low data rates and needs to be deployed remotely, then WiFi is not an option, but likes of cellular, LORA or Sigfox type of network.
So in the long run, there will be multiple technologies in play. Sometimes even working in tandem to provide connectivity. For example, agricultural sensors can be LORA based, and cover large fields, but the back-haul to the LORA network could be 4G-LTE.
Having said that, within in the wide area connectivity technologies, NB-IOT/LPWA stands a very good chance of gaining huge traction, simply because it would be rolled out be Telcos, say Airtel or Vodafone, piggybacking on existing their cellular tower infrastructure that is already in place.
Q. How’s the development of standards (or lack of it) affecting the adoption of IoT?
Considering IOT as a single piece of technology is an oversimplification of what it actually is. It is a combination of myriad blocks of technologies that works like cogs of a wheel. I believe we are still in the early stages of evolution of standards in each of these areas. As an analogy, we are probably still at a 1G stage of early mobile communications. There are areas where there are reasonably well-established standards in place, for example, industrial interfaces and in others, there is a multitude of competing standards, each trying for a dominant space.
However, looking at this from India perspective, there are very little standards that are defined in the IoT space here in India. Local standards bodies are almost non-existent in India and there is hardly any Indian presence in global standards groups which could essentially mean that India is still a consumer of technology and not a creator. The growth of adoption of IoT is because of the new capabilities it brings to table. Standards will follow suit.
Q. How ready is India’s tech eco-system to develop and deploy IoT solutions?
I think India is in a very strong position with respect to developing IoT solutions. It has a vast pool of engineering talent, that is essential for designing and developing complex hardware/software solutions. It is also a center place to find problems and validate solutions given its huge size and growing economic activity. Cost of connectivity is also relatively lower.
However, compared to China, it is still in nascent stages with respect to the manufacturing eco-system. China is vastly superior in the manufacturing eco-system, needless to say, because it is the factory of the world and hence most of the electronic components are manufactured here.
The second challenge is the adoption of new technology, India is slow in adoption of new technologies especially from home-grown start-ups. But once it catches on, the market size is just too big to ignore.
On the brighter side, IOT enables Indian industries especially MSMEs which have missed SCADA like systems phase, jump a generation to Industry 4.0 that is cost-effective, scalable and data-centric.
Q. Do you foresee India’s tech industry developing its own IP and branded products/solutions in the IoT arena?
Absolutely, yes. And that is something that is already happening. There are plenty of companies, large and start-ups that are building IOT enabled products.
But there are areas where Indian companies, Indian start-ups aren’t active yet, especially in areas like chip design or radio technologies, efforts the Shakti Processor from IIT Madras are exceptions. The engineering talent pool is available and a large market base is available but rest of the eco-system is still missing.
Q. Do you see the Open Source phenomenon play an important role in the IoT arena?
Open source could fall in multiple categories: Industry consortium like AllJoyn or OIC, Protocols like MQTT, COAP, OS like ARM embed or Snappy Ubuntu Core, hardware like Arduino, openpicus, Horizontal/Vertical platform like DeviceHive, Eclipse SmartHome, Visualization/Dashboard like Freeboard, Kibana, etc.
Not all projects are full open-source collaborations and come with varying degrees of openness. While there is no one single open-source end-to-end solution, they provide various key building blocks. And that accelerates the development and deployment process of IOT solutions.
Q. How do you see the role, technologies like AI/ML will play in the evolution of IoT solutions?
AI/ML has a key role to play in IoT solutions. While IOT platform gathers so much of data generated by the sensors/devices, it is of no use if you cannot use the data for the decision-making process. And it will be humanly impossible to assimilate this volume of data to convert it to an actionable intelligence. But without this, the data is pretty much useless. And that is where AI/ML plays a key role.
AI/ML processes this data to monitor what’s happening, determine efficiency, identify anomalies, plan for maintenance even before breakdowns happen, predict operational characteristics and so on. For example, here is something that we are building, take a simple use-case where you are monitoring your power consumption with an IOT based monitoring system, and you can receive minute-by-minute information of how much you consume on your mobile phone.
Yes, now I can get a lot more data instead of a bill at the end of the month. However, wouldn’t it be more compelling if the app can tell me how I can reduce and optimize my consumption, if my Air-conditioner is efficient in cooling, if it faulty, if it requires maintenance, if it can predict what will be my next month’s bill, if I can set a target consumption and automatically my consumption gets managed accordingly? It is the same data, but now there is a lot of actionable intelligence. And that will be provided by AI/ML.
Q. What’s your opinion on the state of security available for IoT solutions? How do you the evolutions from hereon w.r.t threats and counter-measures?
Security probably isn’t getting the attention it should. But it is also extremely challenging not just given the complexity of numerous pieces that make up anIoTT solution but also the countless ways it can be deployed, the physical spread.
Take for example, in a traditional network, Firewall, network access control – wired/wireless and physical access control should pretty much take care of security. Throw into this a bunch of wireless light bulbs…and a malicious OTA firmware upgrade, it is not just a case where some unauthorized person can now control the light bulb but is a peephole to the entire network.
And we’ve recently read about a voice assistant device allegedly capturing and sending out conversations in a home or the baby monitor camera being hacked.
While traditional methods of security will help to a certain extent if properly planned and implemented. However, the advent of IoT has opened up a whole new dimension like IOT endpoint protection, network carrier protection, and cloud service provider protection that a traditional security system isn’t capable of handling. Moreover, these themselves may be untrusted and hence the security system has to handle protection of and from these sources.
Security systems employing AI is growing but is still at nascent stages and will gradually evolve to provide robust security.
Q. W.r.t. edge vs cloud–where do you think will we see faster development in the next year or two?
The key aspect of IOT is the ability to collect and push sensor data over the network with very little CPU power. So even a host-less capability provided by a wireless chip which provides CPU power that is sufficient to gather sensor information, do a bit of processing and push to cloud all without the need for a standalone MCU or a CPU.
But that in-effect pushed the entire post-processing of data to the cloud. Which may be okay if the data is limited and use case is tolerant to latency and occasional connectivity loses. However, in use-cases where for time-sensitive applications, the latency has to be very low, large raw sensor data generated with near zero tolerance to connectivity loses, Edge comes to play.
The Edge computing depending on the capability and complexity can either do the entire processing or first level of processing before the processed data is sent out to cloud for further processing. This is critical, say, a fire management system, where the response to the sensor information has to be instantaneous rather than wait for action like shutting off valves to be pushed from the cloud.
Or in the case of machinery that requires processing of huge data generated at milli or microsecond rate, Edge processing will suitable. Now, one could ask, isn’t the on-premise processing already available, in systems like SCADA or industrial PCs for many, many years? Yes, but the key distinction here is that the Edge device is not a full-fledged computer/server, but still an embedded device or SBC, like a gateway or even a network router.
Now, this also brings about another interesting ability to be able to put together IOT and AI/ML on a single Edge device, which now means data collection and assimilation can be done in one device. That is an area that will see a lot of development but will be a niche area.