We detect you are using an unsupported browser. For the best experience, please visit the site using Chrome, Firefox, Safari, or Edge. X

Edge Computing and IoT Products of Tomorrow

The demand for analyzing data at the edge is a growing product requirement. Computing at the edge device instead of the cloud means that…. 

Customers who have struggled to design IoT edge computing products will now be able to understand all of the requirements needed for secure connectivity, data processing, system design and life cycle management, thanks to our total system solution SOM approach. 

Big data, smart devices, cloud connectivity, artificial intelligence (AI), machine learning (ML) and security. These are the terms you hear at every technology conference in existence today. We are swimming in an ever-expanding sea of data, and IoT devices are trying to keep up by analyzing as much as possible in the cloud, while still fighting off security threats and vulnerabilities! A typical IoT-based device is often reliant on a stable and secure internet connection in order to compute data points locally in the cloud, sometimes before it can perform vital task(s).  

So, what happens when that device connection is not stable, or the cloud environment experiences technical issues?   

In this scenario the device might not be able to perform appropriate tasks based on the lack of instruction from the cloud. Now imagine a similar device that can compute the same data points and is able to act without an internet connection, all the while maintaining security at the hardware level. Many devices today are trending in this direction and for very good reasons. 

Due to security risks and customer demands for more intelligence at the edge, an evolution of the IoT market is taking place. Catering to edge technologies, the computational power needed for data gathering, processing, communication and running complex security algorithms is exponentially rising.  All these requirements, when combined with the shrinking size of end products, present some serious circuit design challenges, including system power management.  

While increasing edge computing the remote connection to the cloud remains valuable for collecting summary data from the edge device or over the air updates. 

This article aims to explain these challenges while offering solutions for current and future product designs. 

An example of how IoT needs are changing is illustrated in figure 1 and 2. 

Figure 1. The challenges of intelligence and shrinking solution size.

Figure 2. The needs of customers are changing as well.

Let’s tackle the intersection of intelligence and size first, as seen in figure 1. 

Understanding Computing Power Differences  

Applications designed for edge computing also need to realize the key differences in requirements when switching from a typical microcontroller (MCU) and utilizing a much more capable microprocessor unit (MPU). 

Figure 3 outlines some of the key feature differences. Go here to view a full-size image of Figure 3.

MPU Based Edge Computing Solutions 

Microchip’s MPUs are a natural extension of our 32-bit microcontroller portfolio and may be a fit for you if you fall into any of the following categories: 

Have you outgrown the capabilities of an MCU? 

Do you have the need for large external memory? 

Do you want to a use Linux operating system (OS) which also include the middleware and drivers you need, free of charge? 

Transitioning to an MPU can be greatly simplified by using an MPU-based System-On-Module (SOM) or System-in-Package (SiP) solution. These offerings take away the complexity of high-speed PCB design and provide a robust and reliable power management implementation. Effortlessly move through the hardware development process by using the available development boards, MPLAB® X Integrated Development Environment (IDE) and MPLAB Harmony v3 peripheral libraries and device drivers. 

For those of you using Linux, all of Microchip’s Linux development code is mainlined in the Linux communities providing you full access to source code and long-term support. This also opens the door to the world’s largest developer community providing access to a plethora of qualified software engineers who can assist with your projects. 

Securing the Edge 

Security is a multi-layer strategy and architecture but also a logistic challenge. At the embedded electronic level for IoT systems, the first vital few to have in your system are 1) the implementation of both secure boot to verify the code the system starts from is genuine and 2) the private key linked to the certificate chain to ensure a secure and trusted authentication is not tampered with. The SAMA5D2 will natively support the secure boot and the ATECC608A Trust&GO will make sure the private keys necessary for the device to cloud authentication are protected with physical anti-tampering and side channel attack protections. To further expand the ease of use of the wireless SOM, the private keys are pre-provisioned into the ATECC608A leveraging Microchip’s secure key provisioning service.  

This is a very cost-efficient way to solve the challenge of security logistics. You will save the cost of not only hiring a third-party certificate authority company but also save on test development, implementation and maintenance due to a customized programming. As each key is unique to each device, each system becomes unique. Previously, this system “personalization” causes a logistical nightmare for the product owner, as uniqueness adds consequent costs to their supply chain. But Microchip’s provisioning service significantly reduces all those costs and provides a securely pre-provisioned chain of trust even for low volume projects, thanks to the ATECC608A Trust&GO from Microchip’s Trust Platform. 

Connecting the Edge 

Most IoT applications will require connectivity for cloud-based operations and data collected at the edge. This communication can be accomplished with or without wires based on the location of the device and overall product requirements. For robust and reliable wired communications, Microchip’s KSZ8081 is the perfect choice for critical data communications at speeds of up to 100 Mbps. If running Ethernet cables to your application is not possible, or your application is portable, you should consider the ATWILC3000 UA Wi-Fi Module featuring Bluetooth Low Energy (BLE) for provisioning onto Wi-Fi networks using a smartphone, tablet or computer. The ATWILC3000 supports WPA2 security protocols for both private and enterprise security models. The ATWILC3000 UA has been pre-certified for use in North America and Europe making it much easier and less costly to deploy. 

Timing the Edge 

Timing devices are the heartbeat of an edge computing or IoT system by providing accurate refrece clocks to MCU/MPU and other peripheral controllers in the system. Small size and long battery life become increasingly important with the rapid growth of IoT. The DSC6000 family of MEMS oscillators from Microchip enable unprecedented size reduction and battery life by replacing bulky quartz oscillators that have limited frequency options, or MCU internally generated frequencies that lack accuracy and consume I/O pins. The DSC6000 offers small package sizes of 1.6mmx1, 2mm, 1.3mA current consumption and ±20ppm over -40°C to +125°C. The low output drive strength and spread spectrum clock option improve the system EMI performance. 

Powering the Edge 

Power requirements around an MPU centric solution with smart, secure and connected features can be quite complex and challenging. A PMIC based power solution can address the key issues such as power sequencing, optimal integration of highly efficient and configurable regulators thereby effectively reducing solution size and overall BOM cost. In addition to these features, the MCP16502 PMIC extracts the best performance out of the MPUs with multimode operation, fault management and pin configurable memory power (and pin-pin compatible variants) to address different types of memory enabling platform designs. 

Microchip not only provides all the above hardware, but also the software (free Linux distribution) as part of a fully validated solution in Wireless System-On-Module (WLSOM1). This facilitates a risk-free foundation, leading customers to fast time-to-market and high-quality solutions. 

AWS IoT Greengrass 
AWS IoT Greengrass enables edge computing into an AWS infrastructure. By running AWS IoT Greengrass on a Microchip microprocessor, you can link numerous edge nodes to Amazon’s cloud-based web services and concentrate data and services at the edge devices. It reduces unnecessary data relays to the cloud, thereby lowering the overall solution cost and make the system tolerant to unreliable cloud connections. 

Bringing It All Together as a Total System SolutionATSAMA5D27-WLSOM1 

Microchip has developed a total system solution that solves all the major hurdles discussed in this article. It includes the following solutions and is depicted in Figure 4: 

SAMA5D27-LD2G: System-in-Package (SiP): 32-bit MPU + LPDDR2 

MCP16502: Power Management IC (PMIC) 

MIC842: Voltage detector 

SST26VF064BEUI: NOR Flash memory 

KSZ8081: 10/100 Ethernet PHY 

DSC6102/6003: 25 and 24 MHz oscillators 

ATWILC3000: Wi-Fi/Bluetooth module  

Figure 4. ATSAMA5D27-WLSOM1 Block Diagram. Go here to view a full-size image of Figure 4.

Figure 5. SAMA5D27 Wireless SOM1 (WLSOM1) is a small single-sided System-On-Module (SOM). 

Now you can fast forward to your smart, connected and secure future with Microchip, accelerating your IoT development with a total system solution and world-class global design support. To learn more, visit https://www.microchip.com/design-centers/internet-of-things/amazon-web-services/intelligent-gateways.

Prasanna Obala Bhuvanesh, May 19, 2020