Everything is getting smarter. As the Internet of Things (IoT) continues to develop, our world is becoming more connected—from how cars navigate, to automated factory operations, to multiple intelligent systems within the home that bring greater convenience and efficiency to daily life.
Beneath the visible layer of the IoT evolution, there are also some significant infrastructural changes happening in parallel. Most notably, the high volume of data being generated from a vast number of sensors and “smart” devices in the IoT that could overwhelm networks and datacenters. For this reason, a greater amount of processing and data analytics are being pushed to the edge node. The ability to practically accommodate massive amounts of data in the future will be a critical consideration of smart edge node architectures. Many of today’s edge node devices were not designed with IoT in mind.
The lack of data leverage and the pressing need for a more robust analytics stack are impediments to advancing decentralized intelligent products and systems. IoT edge node devices are generating a greater volume of small packets that are not always best handled at the datacenter, which could lead to packet loss and latency within the connected device. Most data volumes today — and into the near future — are exceptionally low by some standards, such as telco and networking. The key will be providing the underpinnings in the edge node device to service the full analytic stack and feed enterprise applications.
Why lower-end MPUs are becoming more popular
In today’s IoT world there are a number of things that are driving the need for greater processing performance in the edge node. For instance, sensors can accumulate information on air quality, traffic or movement and analyze the information to take action. A factory is now remotely monitored and managed for efficiency, all of the edge nodes need to maintain real-time operations and diagnostics.
More data analytics and processing on the edge node requires more robust processor performance, including a bigger memory footprint (such as that provided by a DDR memory base) to run a full OS.
One of the hardware architectural changes that’s already happening is the shift from microcontrollers (MCUs) to microprocessors (MPUs) — and this is largely because MPUs are required to run a fully featured OS, such as Linux, Android, or Win10. This is critical for IoT devices because connectivity to both the network and other infrastructure comes with the OS, and it also offers built-in drivers.
But there are still many challenges to address in terms of optimizing edge processing deployments. With limited or even no human resources dedicated to maintaining the database and other capabilities at the edge, the system must be a hands-off operation with only remote monitoring/control. The hardware footprint will also be limited by necessity. This means that not all use cases will apply, though many still do, such as:
- Homes (but with far less data)
- Many aspects of smart cities
- Hospitals (but only marginally for personal health)
- Cars (edge on board)
- Other transportation modalities (especially planes, trains, and ships)
- Oil and gas
The NXP i.MX 6UL MPU for data analytics and edge processing
The NXP i.MX 6UltraLite (i.MX 6UL) MPU provides the processing performance (up to 696 MHz) and the memory footprint (DDR3L) to help ensure that data can be effectively monitored analyzed and actioned at the edge node. Its architecture was developed with several advantageous features for auto, industrial, and also home IoT applications.
First and foremost, the i.MX 6UL provides security (built into the BSP). Its power efficiency enables lower-temperature operation, less than 200mW that can be battery operated. Its small size is supported by the smallest Linux machine on the market. It also provides scalability by being compatible with the entire i.MX MPU family.
As the IoT continues to grow, this sort of enhanced processing capacity at the edge node is going to make the difference in giving products and systems the autonomous analytical capabilities they’ll need to make meaningful connections and contribute to our daily lives.