Pursuit of Intelligent Things

Billions of Internet of Things (IoT) devices have become a part of our lives today. Their impact on our lives has also increased manifold. While some of these devices can be seen and felt, others are working behind the scenes tirelessly to provide us with a service.

IoT & Me

• In my daily life, I can't live without my Fitbit! It continually senses my heart rate and SPO2 levels, and brings these insights on a small screen on my wrist. I feel very comfortable when the readings are in the acceptable range.
• Let's take the example of a specialised IoT device for Epilepsy detection - Empatica Embrace that senses motion, skin temperature and electrical activity on the skin to predict seizures. Embrace uses Bluetooth Low Energy (BLE) to connect to a nearby phone in order to notify loved ones of any potential seizures. Personally, I felt even more comfortable when I "did not hear from" my daughter's Embrace for months. No news was indeed good news!
• Obviously, these devices are generating unprecedented amounts of data being stored in the Cloud. The possibilities are endless ...

Secure sharing of fitness data with one's insurance company to adjust premiums, sharing seizure data with neurologists for better treatment, improved citizen programs for better living, building better lifestyle products and so on..

Now, let's think of the not-so-obvious scenarios -

• The electricity we consume daily, the almond milk we use to make our "healthy latte", the latest robotic vacuum cleaner that keeps our house clean.
• There are massive machines involved in the background to generate electricity, in packaging the almond milk to perfection and in assembly lines producing batches of smart vacuum cleaners.
• These equipment need to have the maximum uptime to ensure the required throughput/production volume to fulfil consumer demand. Any disruption due to part failures, lack of spares, or unavailability of skilled technicians can result in a significant loss of revenues and also affect long-term customer loyalty.
• Well-utilized assets also end up becoming more Circular in nature as Circular Buisness Models recommend the?Extend-Share-Recycle?approach where an asset’s life can be prolonged as much as possible (e.g. with predictive maintenance), it can be shared with other customers (e.g. telecom tower sharing) and be recycled for other use cases (e.g. using refurbishing & re-manufacturing).

IoT & The Shopfloor

Zooming into the production shopfloor, for decades, Supervisory control and data acquisition (SCADA) systems have been used for connectivity and monitoring. Data received from SCADA systems could be actioned by the Programmable Logic Control (PLC) systems e.g. a machine could be stopped when its operating temperature exceeds a certain threshold - an efficient combination!

Over time, the product complexity increased, technology improved and supply chains were geographically distributed. Scalability and Interoperability of the operational systems became increasingly important!

The Internet of Things (IoT) evolved as the most appropriate solution to address these requirements.

• It is built for higher scalability managing millions of devices, leveraging lightning-fast network speeds offered by WiFi and 5G.
• It supported Lightweight protocols like MQTT, XMPP and CoAP to connect to the Cloud which is stateless and network efficient
• In addition to that, it also supported short-distance protocols like Bluetooth, Zigbee, LoRa between IoT gateways and the sensors

IoT & Industry 4.0

Now, re-imagine the shopfloor with classic devices that could be "bolted-on" with IoT sensors alongside other modern IoT devices capable of communicating with the latest protocols. On top of that, imagine individual factories getting connected to the Cloud making monitoring and action possible from any part of the world. Mass management of devices with security is key to any modern IoT deployment mapping a "distributed yet connected global factory".

The most exciting aspect of the IoT architecture is the ability to infuse Machine Learning models into the vast amount of data captured. For example, capturing the vibration readings of an equipment over time can result in patterns that can in turn, be used to predict failure. Predictive maintenance of key equipment is of utmost value in a production scenario.

This use of sensors and data to optimize manufacturing and design processes using Analytics & Machine Learning is also called "Industry 4.0".

Hyperscaler Platforms like Azure, AWS and Google offer not only compelling choices for IoT platforms but highly capable Machine Learning model development as well. Leading Intelligence Enterprise platforms vendors like SAP also offer options with SAP IoT and ML as part of SAP Business Technology Platform (BTP).

IoT & Edge

This brings us to the interesting question - does everything need to be sent to the Cloud? How do we balance the use of high network bandwidth and the associated cost with the use case?

• This is where IoT in the Cloud and Edge work in tandem. Edge computing by design is as close to the device or sensor e.g. in a factory, or within an oil rig.
• Sometimes, it is the type of data e.g. video analytics or the immediacy e.g. draw-out spares that necessitates the process to be run on the Edge.
• One can envisage a central repository of Machine Learning models that can be continually trained along with a central Policy layer that maintains the Business Rules / Logic. A periodic push of the Models and Policies to the Edge nodes provides maximum efficiency.

IoT in Supply Chain

So far, IoT seems to do a great job of detecting a situation, providing insights and even taking specific actions. But the list of actions depends on how complex the enterprise use case that IoT is being deployed to solve.

For example, if a machine is overheating, beyond stopping it, should a service from a trusted agent be ordered? is the machine up for replacement? should the orders be fulfilled from another factory?

Historically, the most important use cases of IoT have been in Supply Chain. We explored use cases on the factory shopfloor earlier but there are broader use cases in the Warehousing and Transportation of the finished products as well. The questions above can be best answered by plugging into the broader supply chain context e.g. Procurement Networks, Asset Network, Manufacturing Execution etc.

The Case for IoT

Lastly and most importantly, running an IoT project without a business case is like buying Fitbit with no plans of improving one's health (just that Fitbit costs much much lesser than an IoT project). "Piloting" IoT as a technical project often ends up with an isolated use case with singular/limited business benefits. Most "pilot" IoT projects end up measuring a characteristic and displaying the results in a nice dashboard.

The following questions can help navigate towards a clear value statement and help address the "IoT Pilot Conundrum"

• Why are we measuring this data?
• How is the measurement & related use cases contribute to the top-level KPIs?
• Do we have enough data to make predictions?
• How can we act on the insight/prediction derived?
• Are there broader use cases that we can impact?

Beyond immediate business value, IoT can completely transform business models in some cases. This is due to the inherent capabilities of "measurement" - if something can be measured, an action can be taken; similarly, billing on consumption can also be done. There are several use cases where companies have gone from selling air compressors to compressed air as-a-service and from selling soap to cleaning bathrooms as-a-service.

In Closing

Circling back to the personal example I started this conversation..

Embrace is a personal IoT investment that collects skin and movement data, displays past seizure reports, and communicates via Bluetooth but its real value is in the detection & notification of that single seizure event my loved ones cannot afford to endure!

Note: I specialize in Industry 4.0 and IoT. I work closely with SAP's solution experts and key partners in this area. Please reach out if you would like to have conversations on this topic.