In a casual conversation, Zach, Head of Competency Lab (Cloud, Mobility, data analytics and IoT Solutions),Embitel Technologies India Pvt Ltd, talks to Priya Ravindran of EFY, about the advancements in technology that are propelling the Internet of Things (IoT) forward.
Q. Over the past one year, what are the latest technological trends that have influenced IoT?
A. There has been rapid evolution happening in the area of sensor networks. Digital evolution has a huge influence on where we are going with IoT. From just a couple of standards initially, we now have efficient short-range data transfer via Bluetooth LE, energy harvesting solutions from EnOcean, ZigBee, 6loWPan and so on. New standards are also being formulated that help converge technologies. With sensors getting deployed in almost every segment, expectations have risen wherein businesses have reference points and can better understand the benefits gained in incorporating the latest technology into their products and services.
Q. What are your thoughts on using energy-harvesting sensors?
A. In an age where wearables are ruling the roost, energy-harvesting sensors are making battery-less devices a reality. Energy harvesting sensors use motion, heat and light for charging themselves. They also save on charge by waking up and transmitting data only when required and sleeping at other times. They enable gathering of data and wireless operation even in on-the field situations. Such sensors are a good fit in designing IoT solutions.
Q. How have sensor networks changed over time?
A. Over the last decade, the bandwidth of data transfer has improved and data loss issues have been greatly rectified. Integration with gateways has become more plausible now. Gateways are being custom-designed for developing applications for specific domains (Smart Homes, Connected Car, Industrial, Energy sector and similar others). We can now easily connect a sensor network to a gateway and scale it up to the Internet, enabling seamless communication.
Q. What are the things to keep in mind when designing a gateway?
A. Off-the-shelf gateways provide a narrow feature set and boundaries within which they operate effectively. Extending them to keep up with new technologies becomes complicated. While custom-creating a gateway, there must be certain design considerations taken into account for future scalability. Creating a modular architecture to support both hardware and software enhancements is quite essential. Also, re-usability of modules is good to consider when planning an upgraded version. A custom middleware and board support package (BSP) must still have the element of portability.
Understanding what level of processing and storage is required by the gateway, the data exchange expectations, protocols and the operational scenarios, both in the short and long term, helps during hardware and software design.
Q. How would you go about choosing protocols for a particular design?
A. Since most of the problems or use-case scenarios are addressed by specific protocols, it usually is not complicated to choose the right protocol for an IoT solution. Yet, there are situations where data may need to be made available in more than just one protocol, thereby enhancing usability and reach. While selecting protocols, we need to take into account factors such as bandwidth, latency, data size, security and scalability, and understand the real benefits. Let us say for a retail shop outlet, where the intention is to engage with a customer via a mobile application, you might consider doing away with ZigBee and designing a solution using iBeacons and Bluetooth.
Similarly in the case of an automobile, in-vehicle infotainment requires high-bandwidth data wherein the gateway or the head unit may need to bridge multiple protocols such as controller area network (CAN), Bluetooth, Wi-Fi, universal serial bus (USB) and possibly even others to achieve the desired functionality and connectivity.
Q. Would you have worked on any project where issues were faced due to a wrong choice?
A. It would only be surprising if we do not face any issues. More than a wrong choice, it is the resultant learning that matters. While working on a home automation project, we got communicating with the gateway and realised that it was taking about seven to eight seconds for the lights to turn on, after a command was issued. These commands were sent over Web sockets, from mobile devices. Although it was working fine, we needed much lower latency. We noticed that the delay was majorly because of the time needed to perform socket operation, send data and create the packet format.
We then shifted to an entirely new architecture, doing away with the sockets. We explored protocols that can publish and subscribe, and narrowed down to the IoT-specific Message Queue Telemetry Transport (MQTT) protocol. These support small data packets and provide low latency in communication, and it enabled us to control the lights within a second. In fact, it is quite important to be flexible and adapt quickly to new ways of doing things when working on IoT-based projects.
Q. Having worked on a lot of automation-based projects, what would you say are the latest developments in this field?
A. Machine-to-Machine communication will be a leap towards creating factories of the future. Manufacturing is gradually becoming semi-automated to fully-automated, with robots assisting humans on the production floor. Robots are even capable of doing more than just one job which offers huge benefits and return-on-investment (ROI).
Ability to monitor and optimise workflows constantly as per the needs of the business will be an area where IoT will add value. Also the ability to perform predictive maintenance greatly improves operational efficiency and reduces downtime, which is most necessary. Existing machinery can even be retrofitted with sensors that help monitor and provide essential data using data analytics, useful for expanding to a wide range of services.
Q. What are the challenges in monitoring and securing data from a hardware design perspective?
A. Earlier, we used to mainly deal with static data, getting files over the Internet and performing analysis later. But now, the challenge is the for monitoring and  data analytics in real-time, which depends on factors like the bandwidth, network speed and the amount of data that needs to be operated on. For a seamless user-experience, many systems and components need to be part of the system design, and they should be capable of handling the load and also work hand-in-hand with each other.
Coming to security, it is always a big challenge in collaborative systems. This can be implemented at multiple levels or stages, starting from the sensor level where a decision on what data to send can be made, the gateway where we decide what should be stored or encrypted, and likewise in the cloud. There are lot of secure frameworks (both free and paid products) that cater to the IoT communication and security workflow needs.
Basic measures such as multi-level authentication-based user-access using key-based access and other methods can always be put in place. Yet, at certain times, security can also become an overload, with unnecessary security measures introducing bottlenecks for the flow of data .
