The Road from M2M to IoT is Paved with Big Data

Although machine-to-machine (M2M) communications and the Internet of Things (IoT) have large numbers of connected devices in common, there’s an essential difference between the two: how the data that is generated by all these devices is being used to create value. The IoT is fueled by the convergence of M2M communications and data analytics.

In an older post, “Through the Looking Glass: about the Transformational Power of IoT in the Enterprise”, I wrote about the significant impact of the IoT in enterprises and vertical markets. While traditional M2M applications often target the automation and streamlining of industrial systems and processes, the IoT is a game changer that provides companies with new means for renewing products and services.

Today, we are basically still in a M2M era, with connected devices and communication technologies taking a foreground role, but things are moving fast:

• We are witnessing an explosive growth of connected devices. Unlike in the ‘old’ M2M scenario, where many of these endpoints were existing machines, extended with a SIM-enabled communication module (and often still communicating via SMS or GPRS), the IoT landscape shows a broad mix of SIM, eSIM and non-SIM enabled devices, with or without a built-in IP stack. They communicate over licensed as well as unlicensed spectrum, using a variety of broadband (LTE), LPWA (NB-IoT, LoRa), and short range (Wi-Fi, Bluetooth, Zigbee, Z-Wave) connectivity options;
• M2M solution were mainly designed to monitor hundreds to thousands of remote assets, while IoT networks and platforms will need to scale up to collect data from, manage, and actuate hundred thousands to millions of connected devices;
• From a business perspective, M2M is mainly internally focused and driven by cost savings and process optimization, while IoT applications enable growth, innovation, and an enhanced customer experience;
• Where M2M addresses ‘vertical’ applications, developed as a single purpose ‘silos’ or ‘stovepipes’, IoT is evolving to a ‘horizontal’ model in which devices and applications share functionality and data to support multiple use cases, through a common platform infrastructure.
• Data exchanged between machines is evolving from proprietary, application-specific, structured alerts, alarms and location records towards a mixture of structured and unstructured data (including video, as you will read further down this post) that can be accessed by and shared across multiple applications;
• As storage and processor hardware further evolve, and the capabilities of data analytics and Artificial Intelligence software develop, consumers and businesses will get more and more value from data. From sensing, collecting and monitoring raw data, up to using analytics to create knowledge, and eventually making machines take autonomous decisions;
• The security risks associated with M2M applications are rather limited, as endpoints often communicate over point-to-point (virtual) private network connections. With the massive deployment of vulnerable IoT devices and connected gadgets, and the growing volumes of non-encrypted data exchanged over non-protected networks (like the public internet), IoT security has become an issue of high concern;
• In M2M, one single vendor or SI may take responsibility for design, development and integration of a complete solution. As end-to-end IoT solutions are more complex, their implementation requires a broad scope of technologies ranging from connected sensors and modules, over communication networks, to data analytics software. And many skills, such as network planning and optimization, application development, project management and system integration. As such a collaboration between different stakeholders is needed to speed up the development of new applications;
• Application developers need to move from a ‘brick and mortar’ approach that requires deep knowledge on all underlying technologies, towards a process that puts emphasis on application functionality and value for the end-customer. To realize an ‘any device, any network, any application’ Internet of Things, software platforms are needed that make abstraction of the underlying network and device technology, and provide developers with ‘off the shelf’ enabling functions.

In digital transformation, data is an immensely valuable and often untapped asset. It is said that “big data is the new oil” because, like crude oil, the more the data is refined, the more you can do with it, and the more its value increases (as shown in the image below). Just think of the price of a barrel of crude petroleum compared to one filled with jet fuel….

As an example of how IoT applications can benefit from data, let’s zoom in on the video analytics solution that Nokia recently launched on its IMPACT IoT platform.

According to research firm IHS, there were 245 million professionally installed video surveillance cameras active and operational globally in 2014, of which over 20 percent were estimated to have been network cameras. Deployed in city centers, in public spaces, and along our highways, they stream valuable information to assure public safety and inspect traffic, 24 hours a day and 7 days a week. New cameras installed in 2015 generated 566 petabytes of video data per day, which is equal to all of Netflix’s current users streaming 1.2 hours of ultra-high definition content simultaneously! It is expected that these numbers will keep growing, driven by increasing public safety focus, heightened terrorism threats, and deteriorating traffic congestion.

Today, it’s virtually impossible for e.g. cities to manually monitor thousands of CCTV cameras simultaneously. It would require an insanely huge number of people sitting behind video screens all day, looking at images of traffic and big crowds. This is why intelligent video analytics solutions will be needed, that build upon situational awareness, apply machine learning techniques to understand what’s normal and what’s not, and run innovative applications to deal with specific problems in many different domains.

Nokia’s solution, developed in close collaboration with Nokia Bell Labs, features machine learning-powered video analytics that automatically detect patterns and anomalies in video feeds in real time, such as traffic accidents, speeding vehicles, or unauthorized entry into secure locations, and triggers alerts for further action. There’s no need to transmit full-resolution video over a network or employing an army of human operators to view it. Without even recognizing objects such as pedestrians or cars, the visual algorithms can detect movement in an area through pixel patterns. Over time, they can learn the typical patterns and detect movements that are out of the ordinary.

In conclusion, moving from M2M to IoT holds the promise of creating significant value. Service providers, application developers and end-users will reap the benefits from being able to analyse and reuse data in new applications and business models. As an industry we are ready to start architecting and deploying the infrastructure that takes advantage of economies of scale, and to help laying the road from bespoke M2M use cases to an IoT-enabled programmable world.

Read more about the Nokia IMPACT IoT Platform and our video analytics solution.

Join us at the Mobile World Congress 2017. On the Nokia booth we showcase 12 global use cases which demonstrate our open collaboration with more than 50 companies in our IoT Community ecosystem – enabling innovative business models for transportation, energy, smart cities, and healthcare.

(this article was originally published on the Nokia Blog on 16 February 2017)