Sustaining the IoT—What Will This Take?
February 22, 2018
|By: Kristopher Ardis
Executive Director, Micros & Security Business Unit, Maxim Integrated
Cisco projects that we’ll have 50 billion connected devices by 2020, while Arm is gearing up for one trillion internet of things (IoT) devices by 2035. We probably have the silicon, networks, and manufacturing capacity to support this. But can we sustain this many connected products?
We may not be able to cleanly dispose of billions of batteries each year. Or have a place for all of the discarded devices that stopped working after a year. Or have enough insurance to cover the damages stemming from malicious bot attacks. Heck, we may not even have enough engineering capacity to build all of these different IoT devices within our lifetime.
The greatest promise of the IoT is the ability to invisibly connect things that will ultimately enhance efficiency in our work and in our lives. We've now got the technology to sense a lot of things, along with the ability to process a lot of that data. But does that mean we should infuse everything with connectivity and intelligence? Clearly, the return-on-investment for deploying devices that aren't robust, long-lasting, or highly valuable will be low. To this end, semiconductor companies like Maxim are working to close the "should we build it" gap by providing power-efficient, robust, and secure silicon for sensing and controlling the world around us.
Let's Be Responsible About the IoT
Do you have your own graveyard of electronic devices? Like old phones and laptops, each with their own batteries and obsolete electronic components? What about things like old thermostats and fitness trackers? Many of these devices probably end up in landfills. Now, consider a world with 50 billion or even one trillion connected devices. How will we responsibly manage all of these things? Rechargeable batteries are one option, but even these eventually need to be discarded.
Figure 1. Are you still using your fitness tracker?
This is where semiconductor companies have an opportunity. By extending the life and usefulness of remote sensors, semiconductor companies can help reduce the number of discarded batteries, thus ensuring that the IoT is powered responsibly. Microcontrollers like the MAX32630, for example, have plenty of horsepower and memory space to tackle complex, value-added IoT applications while only consuming 3.4 mW from the battery (backing up 512KB SRAM). Power management ICs such as the MAX14745 wearable charge-management solution ensure that power is delivered efficiently from the battery to the system, minimizing loss from power regulation. And power-conscious sensors made for wearable applications minimize the battery impact of detecting the real world.
With these types of solutions, sensors can last longer and do more processing. Nodes can make more intelligent remote decisions, reducing network congestion when they don’t have to go to the cloud to make decisions. This would further save power by eliminating useless network traffic.
Field Updates for Longer Lasting IoT Devices
The latest connected light switches and refrigerators are great—until someone releases a new home connectivity API, or patches a critical bug in firmware. Whenever updates require human intervention, however, the devices simply become more difficult to manage. Yet, updates are crucial. There may be new players in the cloud computing space to which data must be sent, new functionality, or security risks to mitigate.
Ideally, IoT devices should be able to be updated in the field. Sufficient memory is needed to handle a firmware upgrade. You’ll generally need to be able to store both the old and the new revision of firmware until you’re sure that you have downloaded the correct new firmware. Compared to other low-power microcontrollers, Maxim microcontrollers have more flash and SRAM. They can also extend their address space to external memory chips connected via Quad-SPI, providing the ultimate assurance that embedded applications can upgrade to any degree of complexity to continue to provide value.
That Important Security Matter
Whether we should build the IoT is becoming more clear, especially as we continue to find ways to create devices with low enough power to mitigate the number of batteries and devices that end up in landfills. We’re also able to build devices that are flexible enough to be upgraded for future needs: new home connectivity protocols, new functionality, or just fixing bugs in what got deployed in the first place. What continues to loom over the IoT is the matter of trust. Our smart things are gathering intimate data about us: they know when we are home, what we are watching on TV, and what appliances consume electricity in our house, for example. There are also those IoT devices that control critical functions in our homes: can electricity be delivered to this house, what will the temperature setting for this house be, is the garage door down or up? Then there’s the industrial IoT, where threats could be even more damaging: will your city continue to have reliable electricity, will nuclear power plants operate as intended, will manufacturing lines operate productively or will they break due to malware?
Bottom line, we need to be able to trust the IoT—and it must continue to remain invisible. These goals may seem to be at odds with one another. After all, maintaining trust means constant security patches, which require user intervention (and, thus, ruin invisibility). Fortunately, there are techniques to design security into cost-effective IoT nodes right from the start. Today’s low-power microcontrollers and companion processors have security capabilities that let you create a Bluetooth heart-rate sensor equipped with the same cryptographic tools as a credit card terminal. IoT nodes can take advantage of advanced techniques such as elliptic-curve cryptography, making systems widely deployable without risking system secret keys. By using special hardware to implement these cryptographic techniques, the IoT nodes can perform their secure functions more quickly and with less power consumption than if they tried to do so in software.
Figure 2. A low-power Arm® Cortex®-M4 with FPU processor, the MAX32631 also integrates high-performance cryptographic engines.
If IoT nodes can boot securely only from trusted code bases, this sets the stage for future, invisible feature upgrades and even security patches.
Should We Build It?
While we’ve got the technology to build plenty of IoT nodes, we also must solve some real-world challenges before the ROI calculations fall in favor of deployment. We must mitigate the battery replacement issue and extend product life to maximize the potential lifetime of sensors and actuators. IoT devices must be built to be able to update as communications protocols and processing demands change. And we must be able to trust the data we receive along with the commands we send.
The goal is to make the IoT trusted as well as invisible. To this end, semiconductor companies like Maxim are working to make sure IoT devices get deployed once and get forgotten.