IoT devices require various hardware building blocks, and some of the critical ones are processing chips. The primary processing chips are MCUs, SoCs, and MPUs, all of which have different properties that suit various applications. Let’s differentiate these three to highlight their strengths, weaknesses, and factors to consider when choosing one.
What Is an MCU?
A microcontroller unit (MCU) is a compact integrated circuit that houses the microprocessor, memory, peripheral interfaces, and I/O in the same chip. These ICs are not as powerful as SoCs or MPUs, so they are ideal for small, embedded systems, including some IoT gateways.
The primary advantage of MCUs is they consume very little power, which makes them ideal for battery-powered IoT end nodes, such as sensors, which can run for over a year without requiring maintenance.
Advantages of MCUs
- Low power consumption (they operate at slow speeds)
- Cost-effective (cheaper than MPUs and SoCs)
- Small size (suitable for applications with limited spaces)
Disadvantages of MCUs
- Low processing power
- Limited memory capacity
- Cannot run operating systems
Applications of MCUs
- IoT end nodes and some gateways
- Robotics, drones, and automotive systems
- Medical devices (blood glucose monitors, heart rate monitors, and blood pressure monitors)
- Home appliances (washing machines, microwaves, blenders, fridges, etc.)
What Is an SoC?
SoC is an acronym for System-on-a-Chip. As the name suggests, this integrated circuit contains and integrates all the components of a computer or electronic system onto a single chip. These components include a microprocessor, memory, and I/O interfaces. It can also have a Digital Signal Processor (DSP), GPU, and Application-Specific Integrated Circuit (ASIC).
These chips pack more processing power than MCUs but don’t reach the level of MPUs. Regarding power consumption, SoCs draw more energy than MCUs, but not as high as MPUs.
This balanced performance makes them ideal for battery-powered devices, such as tablets, phones, and IoT gateways.
Advantages of SoCs
- Balanced performance (high processing power and low power consumption)
- Integrates multiple peripheral devices, which can reduce the size and cost of the end product
- Optimizable for different applications, such as IoT, automotive, and mobile device use
- Simplifies product design and development because all components are on a single chip
Disadvantages of SoCs
- Not as modular as MPUs
Applications of SoCs
- IoT gateways, especially those with edge computing capabilities
- Mobile phones, tablets, and other powerful portable devices
What Is an MPU?
A Microprocessor Unit (MPU) is a single integrated circuit that focuses more on having the core functions or elements of a computer on one chip. These include the central processing unit, arithmetic and logic unit, floating point unit, clock, I/O interfaces, control unit, memory management unit, and cache.
Some MPUs can be specialized to serve specific functions, such as graphics processing, neural processing, and digital signal processing.
Since microprocessor units focus more on processing digital data, they perform calculations faster and execute instructions more rapidly. Therefore, they are ideal for larger, more powerful computing systems, such as servers and PCs.
However, this high processing power means MPUs consume a lot of energy. Also, since the chips focus on performing the core functions of a computer, they don’t have internal memory for storing programs and data. This data is usually kept in external non-volatile memory and is loaded to an external volatile memory (DRAM) during startup. The external data and memory access makes MPUs startup slower than MCUs and SoCs.
But there is the flexibility of increasing the peripherals like memory to any size that suits the application because this chip system is highly modular.
Advantages of MPUs
- High processing power
- Can run a wide variety of functions or applications
- Highly modular computing chip
Disadvantages of MPUs
- High power consumption
- Expensive
- Overheating (requires active, passive, or both cooling systems)
- Large size
- No internal peripherals like RAM and ROM.
- Slow startup speed
Applications of MPUs
- Personal computers
- Workstations for specific tasks, such as video editing and computer-aided design
- Servers
- Gaming consoles
MCU vs. SoC vs. MPU: Factors To Consider When Selecting One
| MCU | SoC | MPU | |
| Cost | Low | Medium | High |
| Processing Power | Low | Medium | High |
| Power Consumption | Low | Medium | High |
| Size | Small | Medium | Large |
| Memory Capacity | Low | High | External |
| Peripherals | Medium | Many | Mostly external |
| Modularity | Low | Low | High |
| Simplicity | Simple | Medium-complex | Complex |
| Security | High | Low | Low |
In a simple word, the MCU integrates on-chip peripheral devices, and you can directly add simple peripheral devices (resistors, capacitors) to run the code. The MPU does not have peripheral devices (such as memory arrays), it is a highly integrated general-purpose processor, and it is an MCU that removes integrated peripherals. An MCU is chip-level, while an SOC is system-level. SoCs combines the advantages of MCUs and MPUs, having built-in RAM and ROM while also being powerful like an MPU. It can store and execute system-level code, meaning it can run operating systems (primarily Linux OS).
Which Chip Do IoT Sub-Devices Need?
IoT sub-devices include components like IoT sensors, actuators, panic buttons, and beacons. These devices only perform a few tasks, such as sensing changes in vibrations or smoke levels. Therefore, it makes more sense to use microcontroller units as the brains behind their operation.
For instance, a typical sensor requires an MCU to check if the readings meet a certain threshold to send a specific signal.
In a smart home, for example, a temperature and humidity sensor can send an HVAC activation signal to the hub if the sensed temperature exceeds a certain preset value. Alternatively, it can send the raw values to the edge computing gateway, which will determine whether to activate the cooling system.
Why Use MCUs for IoT Sub-Devices
Simplicity
Microcontroller units are simple chips compared to SoCs and MPUs, which is an advantage in IoT sub-device applications. These ICs don’t require operating systems to run, which makes them easy to interface with sensors, actuators, motors, etc. Coding them is relatively easy.
Also, they don’t need external dependencies, which makes them easy to set up. All you have to do is power them, upload the firmware, and run the code.
Security
MCUs have minimal avenues of attack because they have few open ports and protocols. Each open port and IoT protocol provides an access channel, so the fewer, the better from a security standpoint.
Additionally, MCU code runs bare metal with no OS in between to execute the instructions. This architecture limits the attack vectors.
Low-Power Consumption
IoT end nodes need low-power chips to reduce maintenance costs. Microcontrollers provide this advantage, working hand-in-hand with low-power communication protocols like ZigBee and BLE to lengthen the battery life in these devices to 1+ years.
Low Cost
Microcontroller units cost as low as $1 or less per piece, making them extremely cost-effective for IoT projects. Despite their low price, they provide the required processing power and functionality in IoT end nodes.
Which Chip Do IoT Gateways/Hubs Need?
IoT gateways can run on MCUs, MPUs, or SoCs. But most run on SoCs because these chips consume less power than MPUs and provide higher processing power than MCUs. They are sort of the midpoint between MCUs and MPUs, which is suitable for this application because modern gateways need edge computing power and energy efficiency.
Some IoT gateways, such as the DSGW-210 edge computing gateway, run on a powerful and efficient Rockchip RK3328 Quad-core Cortex A53 SoC. The efficiency of this SoC enables the gateway to run on a lithium battery power backup if there is a blackout to operate the network.
But some gateways, such as the DSGW-092 ESP32 IoT Gateway, run on MCUs. This specific hub features an ESP32 microcontroller. The chip achieves extremely low power consumption, making it suitable for building energy-efficient IoT networks, such as entry-level smart-building and smart healthcare projects.
DSGW-340 Portable BLE Gateway is a power-efficient BLE LTE/WiFi gateway with compact design for easy carry at home and the outside. It is designed with ARM Cortex-M33-based EFR32BG24 MCU, and equipped with LTE CAT 1 and Wi-Fi 2.4G module, perfect for remote patient monitoring solution and connecting medical wearables like blood glucose meters and pulse oximeters.
MPUs are not as common because they consume a lot of power. Usually, advanced gateways use MPUs for additional functions, such as graphics and neural processing. These can be independent chips or part of the SoC as processing units.
For instance, the DSGW-380 industrial AI edge computing gateway features a Rockchip RK3588 SoC that has a CPU, GPU, and NPU in the same chip. The CPU has a microcontroller for low-power control, as well. Externally, this edge device has a built-in NEON coprocessor that enhances the SoC’s AI processing capabilities.
However, this high processing power means the DSGW-380 consumes relatively more energy. It has an active cooling system (fan) as well to keep the chips operating at their optimal temperature when performing intensive AI computations.
Why Use an MCU for IoT Gateways or Hubs
- Low power consumption (ideal for battery-powered gateways)
- Cost-effective
- Compact size
Why Use an SoC for IoT Gateways or Hubs
- Provides a balance between low power consumption and high performance (can use batteries for backup power)
- Integrates more peripheral devices
- Can run operating systems, hence providing more edge computing functions
Why Use an MPU for IoT Gateways or Hubs
- For extremely high processing power applications like AI computation
- Can run operating systems, hence more edge computing apps/functions
- To enhance the SoC’s functions externally, such as graphics processing on the edge
Final Thoughts: The Future of MCUs, MPUs, and SoCs
Although there is a clear distinction between microcontroller units (MCUs) and microprocessor units (MPUs), SoCs have blurred the line between them. SoCs are basically more powerful microcontrollers that can integrate NPUs, GPUs, and DSPs into the circuit while running on low power.
SoCs have become extremely powerful of late to the point of replacing MPUs in PCs. Case in point is the Apple M3 SoC, which has replaced Intel chips in Mac desktops and laptops.
Therefore, SoCs will be dominant in most computing applications in the future, including IoT gateway edge computing, while microcontrollers will run the end node layer of the network.
If you need custom-built IoT devices using these chips or want to buy ready-made, vertically integrated gateways and end nodes for your IoT solutions, contact us for further details or check out our IoT products.
