Data alert: IoT analytics essentials

Many standalone devices of the past will soon be networked, and sensors embedded in consumer products and industrial machinery will become ubiquitous — the Internet of Things (IoT) era is upon us. According to a recent survey of 300 IT decision-makers conducted by Frost & Sullivan and commissioned by Singtel, nearly 40% of organisations across Australia, Hong Kong, and Singapore have begun their IoT journey¹.

The sheer amount of data generated by IoT devices will entail that an entirely new set of rules need to be formulated to collect, process and analyse that data. The importance of real-time gathering and processing of streaming data becomes critical, as conditions in systems fluctuate and test an organisation’s ability to accurately interpret large volumes of fast-moving data. A primary component of enabling this capability is to move the processing and analysis of the data closer to the IoT device and decentralise analytics by shifting it closer to the edge.

Here are some essentials to help you get the most out of that data.

IoT devices often generate simple streams of data constantly, which necessitates that any analytics platform connected to them needs to continuously query that data. This calls for a new approach to analytics, such as when sensors in large numbers deployed across physical distances generate tremendous amounts of data that needs to be processed and possibly used quickly.

Because of the often rapidly-changing conditions that IoT devices are subjected to, as well as the mission-critical value of the data they generate — in healthcare devices, for instance — the data must not only be processed in near real time, but also closer to the source of its origin. Major performance improvement capabilities are enabled by edge gateways, which act as the bridge between the device and the data centre. This ensures that the gap between origin and compute is minimal, making it more accurate as a source of truth.

The challenge with data when it comes to IoT is not only in the realm of compute, but also in the peripheral storage and retrieval systems, as well as in the intersection of IT expertise with domain expertise in the field that the IoT system is being deployed. The operational efficiency of automated workflows can reduce the pain involved in manually allocating resources — combining analytics with automation can be an essential initial step when it comes to IoT deployments that need a unique blend of vertical-specific knowledge and deep IT skills to be optimally successful.

For example, in industrial systems, the outcome of automated analysis at set intervals across data streaming from scores of devices can be used as the “seed” data for the next cycle of analytics. This can constantly improve the analytics process and enable longitudinal analysis of an IoT system and the setting of accurate KPIs. Detection of anomalies and the sounding of alerts can similarly be automated when built into the analytics layer.

One of the ways that IoT will challenge traditional computing is in the way it will add layers of complexity and a multitude of protocols to the architecture that engineers have struggled for years to simplify. A variety of devices with their own firmware, their own APIs and hardware idiosyncrasies might threaten to overwhelm the analytics engine and slow down data gathering. Standardisation, therefore, remains a key challenge to building an integrated analytics hub that offers seamless interconnection for all the devices within a system.

The data generated by IoT devices is often messy, complex, and unstructured — its sources are often crude analogue devices such as thermometers and photosensors deployed in rugged environments. Traditional analytics tools meant for discrete, digital and structured data may fail to deal with the “fuzziness” of the physical world. Emerging tools such as artificial intelligence (AI) and machine learning can be adopted to minimise the “fuzz” in these systems and separate the signal from the noise.

An enterprise truly maximises the benefits reaped through IoT when it becomes a “Cognitive Enterprise” — one that is able to explore new business models and digital services by leveraging AI to unlock even more value from IoT². A densely interconnected IoT layer, powered by AI tools and advanced analytics, is able to feed a tremendous amount of actionable intelligence to the line of business that can have a profound impact on the bottom line of an enterprise.

Speak to us to discover how to unlock more value from IoT.

¹ “The Internet of Things Actualisation Quotient: An Asia-Pacific Perspective”, Page 18.

² “The Internet of Things Actualisation Quotient: An Asia-Pacific Perspective”, Page 6.