When you imagine the future of IoT, you probably don’t think of cows wearing sensors that collect movement data to reveal fertility information. Or, for that matter, wandering Alzheimer’s patients who can be located by smart city streetlights. If those scenarios don’t come to mind, it’s no surprise.
Already Internet of Things (IoT) technology has turned many assumptions upside down – assumptions about where, how and how quickly we can collect and manage data, analyze it, make predictions and do modeling. According to Timothy Chou, author and lecturer at Stanford, people used to know what problem they were trying to solve before they started collecting data to analyze. With the IoT, that model of discovery is being flipped on its head.
Consider the cow example. Researchers who attached motion sensors to cows learned that just before a cow is fertile, she starts moving in a much different pattern. The optimum time for artificial insemination is 16 hours later. Researchers also learned that insemination performed within the first two hours of that window has a much higher probability of producing a female – exactly what you want for a milk cow.
No one would have figured that out without the motion sensor data. Because no one expected it – they weren’t trying to solve that problem.
A glimpse into the future
Chou explains: “Most technology we’ve built so far was for the Internet of People (IoP). Whether it was an e-commerce, ERP or search application, it was built to serve people – and to accumulate specific types of data that we could analyze later. But people are not things. Things produce vast amounts of data nonstop and can give us information whenever we ask for it. So why would we think the technology we built for the Internet of People would work for the Internet of Things?”
The future of IoT presents an opportunity to connect all sorts of different devices, collect many different types of data, and learn from it without having to sort it all out first. In the future of IoT, we may be able to learn from things like wind turbines, scissor lifts or blood analyzers before we even know what we’re looking for or trying to accomplish.
Data on the edge – the starting point
There’s a continuum of points in the IoT where data can be generated, collected, aggregated, analyzed and stored. While those points vary with each situation, the “edge” is where it all starts.
Everything generating data outside of a data center and connected to the Internet is at the edge,. That includes appliances, machines, automobiles, streetlights, smart devices in the home, turbines, locomotives, pets and healthcare equipment.
Some amount of intelligence and computing power can be placed in edge devices with today’s technology. But data can’t be fully analyzed at the edge yet because most edge devices don’t have sufficient computing and storage resources to perform machine learning and advanced edge analytics. As a result, many IoT applications observe data at the edge, then move it to the cloud for analysis.
In the future of IoT, says Kumar Balasubramanian, General Manager of Internet of Things Solutions at Intel, “Industries that stand to gain the most are those that are able to extract the right business insights at the right time and the right place – edge or cloud – based on factors like cost and latency of the underlying business problem.”
So how do you determine what runs at the edge versus what runs in the cloud? You have to decide based on the situation.
Here’s an example. If a smart car senses that a driver is about to have a stroke, you can’t wait for the data to go to the cloud to be analyzed, then wait for a signal to come back to the edge device to direct the proper action. The cloud is too far away to process the data and respond in a timely manner. As Schabenberger puts it, data has an expiration date – you can’t afford to devalue the data before you apply analytics.
There are other issues with sending raw data to the cloud through the internet, too. Think of the privacy, security and legal implications. Anyone planning for the future of IoT will need to weigh these considerations.
Industries that stand to gain the most are those that are able to extract the right business insights at the right time and the right place – edge or cloud – based on factors like cost and latency of the underlying business problem. Kumar Balasubramanian General Manager of Internet of Things Solutions Intel
Enterprise-class thinking should direct the future of IoT
“Things have changed with the IoT because embedded logic is no longer isolated logic,” says Schabenberger. “It is connected logic. This means we have to touch it. We have to be able to update it.”
Let’s say you find out with analytics in the cloud that a model scoring something on an edge device needs to be updated. In this case, you’ll want to send out a new champion model without disrupting the operation. For the future of IoT, that type of thinking needs to be in every element and every component on the edge.
Schabenberger explains: “It is all about bringing enterprise-class thinking from the cloud to the edge and to everywhere in between. That includes all the components; and it requires baking enterprise thinking into the software within the devices or the things.”
Scalability and agility
Fortunately, cloud computing has been around for a while, and many enterprise-class assets are already built. This includes agile, available and scalable services. As you design for the future of IoT, you should go through every layer and ensure that security models, data models and manageability models are consistent at every stage.
For example, let’s say you try out a new service by starting with 40 edge devices, and it works well. But what happens when you want to go from 40 to 400 devices? If you defined the architecture in a scalable way from the start, you’ll be ready to make that leap. Every element at each layer of the stack has to be poised for reliable, secure and scalable service delivery.
Keep in mind that the nature of the business value you want to extract is going to shift over time. You need to think through your design for that level of scalability upfront.
Along with the IoT comes the need for unprecedented levels of security. This means designing with the future of IoT in mind. Because when you’re out in the real world, security is a very different thing than it is within the walls of your data center.
Think of an IoT infrastructure that helps manage traffic in a city. If there were a security breach in this setting, it could create havoc. To avoid potential security issues, you need to build in security from the start, because it’s hard to do as an after-thought.
But IoT security is complex. For one thing, who owns it? In a connected car environment, for example, security isn’t owned by a single entity. “How do we brokerage security across the multiple parties that own it at different layers and different aspects of the flow?” asks Balasubramanian. “Those are important issues to wrestle.”
In the next 10 years, the IoT revolution will dramatically alter power, water, agriculture, transportation, construction, health care, oil, gas and every other industrial sector of the economy. These sectors account for nearly two-thirds of the global GDP. Timothy Chou Author and Lecturer Stanford
Services at the edge, new business models and partnerships
With more computational horsepower at the endpoints now, networks are starting to connect all sorts of things together and collect the data. With insights from this IoT data, many businesses are already boosting productivity and driving greater operational efficiencies.
In the wind turbine industry, for example, each turbine may have 400 sensors on it collecting terabytes or even petabytes of data. This data helps with predictive machine maintenance and performance optimization. Precision agriculture is another industry where IoT analytics is used. The data that's collected helps farmers operate more precisely, using smaller amounts of fertilizer and herbicides – an approach that’s cost-effective and creates healthier produce.
Streaming data from the IoT helps organizations understand their businesses better because it allows constant monitoring. That uncovers things that couldn’t be seen with just periodic data views. The end result is often a more customer-centric business model. But to get the full value out of the future of IoT – for the business and the customers – you’ll need to have a data-driven culture that allows the data to drive your insights.
The future of IoT presents the opportunity for new business models as well. It opens the door for organizations to become information vendors – not just metal-box type vendors. That changes the strategic approach. In this environment, your business will need to have a product and services mix to succeed. If it doesn’t, you’ll miss an enormous opportunity, and the chance for recurring revenue. “In the next 10 years,” Chou says, “the IoT revolution will dramatically alter manufacturing, power, water, agriculture, transportation, construction, health care, oil, gas and every other industrial sector of the economy. These sectors account for nearly two-thirds of the global GDP.”
Don’t overlook that the future of IoT is based on a complex infrastructure. This will require partnerships among different types of vendors to keep everything working together. Software may be the brains, but it doesn’t account for all aspects of the IoT infrastructure. Hardware, middleware, networking and other components will all need to be compatible in the future of IoT.
Get ready for the possible
The technology for the future of IoT is here. As Chou says, “There’s an enormous opportunity now for a new generation of technologies to emerge to address the very different nature of the IoT. Now we have the chance to learn first. Now we can connect devices, collect a lot of data, learn from it and then figure out what to do. It’s a new way of looking at the world. We’re just at the beginning of what’s possible with the future of IoT.”
- Article Viking transforms its analytics strategy using SAS® Viya® on AzureViking is going all-in on cloud-based analytics to stay competitive and meet customer needs. The retailer's digital transformation are designed to optimize processes and boost customer loyalty and revenue across channels.
- Article IFRS 9 and CECL: The challenges of loss accounting standardsThe loss accounting standards, CECL and IFRS 9, change how credit losses are recognized and reported by financial institutions. Although there are key differences in the standards for CECL (US) and IFRS 9 (international), both require a more forward-looking approach to credit loss estimation.
- Article ModelOps: How to operationalize the model life cycleModelOps is where analytical models are cycled from the data science team to the IT production team in a regular cadence of deployment and updates. In the race to realizing value from AI models, it’s a winning ingredient that only a few companies are using.
- Article Medicaid and benefit fraud in 2018 and beyondTo curb the growing amount of Medicaid and benefit fraud and improper payments, agencies and their commercial counterparts need fraud and abuse detection systems with data management and analysis that can keep up and even stay one step ahead.