IoT, the Internet of Things, is the network of computing devices embedded in everyday objects that enables them to send and receive data. These devices are extremely diverse, depending on the use cases. For example, Amazon’s Echo device takes as input a human voice command “Alexa, play Hawaiian music,” sends it to Amazon Cloud, brings back Hawaiian music and plays it from the device. An industrial IoT device might be a GE engine that has 2,000 sensors in it sending maintenance data to the back end.
Besides IoT, we are seeing a number of technological sea changes in cloud computing, high-speed Long-Term Evolution (LTE) wireless communication networks and artificial intelligence (AI). All these cutting-edge technologies are converging — a major trend that enterprises need to understand when they create IoT strategies.
Internet of Things
According to Gartner, there were eight billion connected “things” (aka IoT devices) as of 2017, and there will be 20 billion of them in 2020. These numbers can be even bigger depending on who you talk to. IoT devices are now not just simple sensors; most of them are actually tiny computers which have compute, storage and communication capabilities. These “smart” devices will bring a plethora of business benefits, but also a whole slew of problems around security and privacy.
In the last few years, we have seen the IoT industry make significant advances. With its ecosystem maturing, we have witnessed pervasive adoption of IoT across many industry verticals. Large enterprises such as Microsoft, Amazon, Google, T-Mobile, AT&T, GE, Boeing and many other companies are all active in the IoT space.
High Speed LTE Wireless Communication Networks
We currently use 4G LTE through our mobile phones. Carriers are now testing 5G LTE and plan to roll out 5G LTE networks in 2020. Compared with 4G, 5G will provide faster speed (1.4 Gbps median speed), higher capacity and lower latency. These advanced features will bring about new use cases described below.
Elastic Cloud Computing
This basically offers “unlimited” computer resources– unlimited compute and unlimited storage capabilities. This means companies can drastically cut IT costs by leveraging cheaper cloud services. They can also gain business flexibility by scaling up or down compute and/or storage resources as needed.
AI went through a long winter for two decades, finally emerging as a powerful technology in the last few years, with Google TensorFlow, Microsoft Azure AI, Amazon Machine Learning and IBM Watson gaining traction in large enterprises.
The above technologies converged nicely in the IoT space, as they form the three layers an IoT solution usually has:
- IoT devices
- Communication “pipe”: an LTE network that provides connectivity between the devices and cloud
- Analytics “engine” in the remote data centers (i.e. cloud), powered by Business Intelligence (BI) and machine learning /artificial intelligence algorithms
What does this technological convergence mean for enterprises? The combined powerful features of IoT, LTE, cloud and AI will take IoT from being simply a means of connecting devices to deliver basic business value like tracking and control, to a disruptor that creates huge business value in edge computing, high bandwidth applications and predictive analytics. Let us examine those IoT layers again:
1. IoT Device Layer:
Smart devices are tiny computers. This allows enterprises to move certain time-critical functionality to the “edge” i.e., the device layer at the edge of the wireless network. That provides high business value in urgent, mission-critical scenarios where you need real-time insights for quick decision making, because you cannot wait for data to get sent to the cloud, analyzed and sent back through the communication pipe. This is particularly useful in manufacturing. Boeing, for example, has implemented IoT edge computing devices in their factories that can stop automatic production processes immediately to prevent possible worker injuries.
In the smart car space, edge computing is also critical. If an autonomous car is approaching a red light, it cannot afford any delay to send the sensor data to the cloud and wait for a control command back. When the car senses a red light, it needs to perform analytics on the edge (i.e., in the car), instead of in the cloud, and stop right away. Intel has rolled out several powerful chips enabling commensurately powerful edge computing. The chips can be programmed to provide insights using machine learning models running on the device, performing both predictive analytics and anomaly detection.
2. LTE Network:
Carriers are offering LTE variations with different power consumption and cost levels, such as high speed/high cost LTE Cat 4 (“cat” is short for category) and lower speed/low cost LTE Cat 1, LTE LTE Cat M1 and LTE Cat NB1.
- LTE Cat 4 is used for our smartphones.
- LTE Cat 1 is a medium speed LTE standard, ideal for a vast number of more feature‑rich M2M and IoT applications, including those that require video streaming and voice support. It has speeds of 10 Mb/s downlink and 5 Mb/s uplink.
- LTE Cat M1 is a low‑power wide‑area (LPWA) air interface that connects IoT devices with medium data rate requirements (375 kb/s upload and download speeds). It enables longer battery life cycles and greater in‑building range than cellular 2G or 3G LTE, or LTE Cat 1.
- LTE Cat NB1, aka Narrowband IoT (NB‑IoT), is another LPWA technology that works virtually anywhere. It connects devices on already established LTE networks, and securely and reliably handles small amounts of infrequent two‑way data. Its data speed is very low at about 200kb/s. It has very low power consumption and excellent extended range in buildings and underground. It is widely used for smart city and energy meter reading use cases.
Industrial enterprises can now pick and choose the best LTE flavors for their use cases. For instance, some companies only need a small amount of data transmitted infrequently (e.g., meters for reading water levels, gas consumption or electricity use). In this case, LTE Cat M1 or LTE Cat NB1 may be the best choices.
The arrival of 5G LTE will be a game changing event as it will offer faster speed, lower latency and lower cost per bit than 4G. This will enable uses cases such as autonomous car communication, and augmented reality (AR) and virtual reality (VR) that require high speed, low latency and low cost. 5G LTE will enable many compelling uses cases and create new market opportunities.
3. AI and Cloud Computing:
This is the brain of IoT and provides the most value in the IoT technology stack. Enterprises can now not only reduce IT costs by moving their IT infrastructure to the cloud, but they can also leverage IoT analytics platforms from Microsoft, Amazon, Google or IBM, to extract economic value out of the IoT data gathered. For example, in the smart building arena, an intelligent IoT platform using machine learning/AI can better manage heating, cooling and room-booking systems by analyzing the usage data on the building.
AI, however, is still in its early adoption stage. We have solved the easier AI problems, such as image pattern recognition, text-to-talk, etc., but the hard problems related to semantics remain to be tackled. For example, having a two-way meaningful dialogue with voice-enabled devices in a random context is still a challenge.
Use Case: T-Mobile’s IoT Service–eSIM (Embedded Subscriber Identity Module) for the Fleet Management Industry
At T-Mobile I led a cross-functional team that implemented a large-scale, global IoT service called eSIM for the fleet management industry. It works like this. A trucking company, for example, that has hundreds of rigs travelling frequently between the U.S. and Canada often needs to track their locations, communicate with the drivers and perform vehicle diagnostics. A single truck might easily use 100 MB of data per month while traveling in Canada. The roaming cost for that truck alone could be as much as $200 per month, or nearly $2,400 per year. T-Mobile’s eSIM card provides a powerful edge computing capability, enabled by backend cloud computing, that can automatically swap the default wireless data plan to a local carrier’s plan without incurring roaming charges–as soon as the truck crosses the Canadian border.
Multiple technologies were used:
- Device layer: the IoT device has T-Mobile’s SIM card inside, provisioned with international data plans.
- Network: T-Mobile’s LTE network and Canadian Wireless Carrier Rogers’ LTE network are used.
- Analytics and control: T-Mobile partners with backend/cloud software providers that provide all the logic in detecting truck location and executing the carrier profile swap. They can also provide machine learning analytics as needed.
Companies can build eSIM into an assortment of connected product categories, including mobile health and wearables, smart connected machines, connected cars and connected medical services among others.
The Convergence of Technologies
In summary, the convergence of IoT, LTE, AI and cloud will help enterprises realize immediate business value to improve operating efficiencies and drive revenues in the short term. Across vertical industries, this convergence will also bring about new, long-term market opportunities that do not exist today. However, like any new wave of technologies, there will also be challenges in many areas in the IoT space, such as device security, privacy and the reliability of AI predictions or recommendations. These problems will be solved as the IoT ecosystem matures.
Dan Mo currently works as a Senior Management Consultant at Starbucks and is leading large-scale, global technology initiatives. Previously, he served as the Chair of the IoT Group at T-Mobile and led a large-scale, global IoT Program eSIM (Embedded Subscriber Identity Module) for the fleet management industry. He is an expert on IoT and artificial intelligence (AI). He did strategy consulting work with many Tier 1 consulting firms including McKinsey and Keystone Strategy. He is an IoT evangelist and practitioner and frequently speaks at industry events. He was the director of the Mobile Ecosystem Forum representing T-Mobile International. In his early career, he held management and technical roles at AT&T and three startup companies. He holds a M.Sc. degree in Artificial Intelligence and an MBA. He can be reached at firstname.lastname@example.org or through LinkedIn.