With the deepening of people’s cognition and practice in the field of Industrial Internet of Things, more and more enterprises have begun to try to migrate computing services from the cloud to network edge processing and execution, which is conducive to avoiding unnecessary network noise and the transmission of massive data. Congestion during wireless network communication is also of great benefit in ensuring real-time and safety in industrial applications. Therefore, industrial edge IoT applications are becoming a hot spot.
In the actual development of industrial edge applications, developers often encounter two “pain points” when making technical decisions for the expansion and deployment of their IoT projects. Paying unnecessary extra costs; second, using an industrial PC with a router to build your own solution, starting from scratch, will inevitably lead to an increase in the failure rate, affecting the progress of the project and future operation and maintenance.
In order to solve these pain points, it is obviously a more sensible choice to adopt an industrial-grade edge computing gateway with high bandwidth, low latency, high reliability and scalability. It can make the deployment and implementation of industrial applications more convenient, and at the same time, in terms of performance and functionality Meet the design requirements. Therefore, in the process of building a powerful and ubiquitous industrial Internet of Things, industrial-grade edge computing gateway products have become just needed.
01 “Core” of Industrial Edge Gateway
Of course, it is not a simple matter to build such an industrial edge computing gateway, especially in the choice of hardware development platform, it must be supported by a powerful processor to meet current and future application requirements .
Developers’ demands for the “core” of industrial edge computing gateways can be summarized in the following aspects:
1 Strong computing power: Provide a sufficiently powerful computing engine for edge processing of data. In more and more applications, it is also necessary to consider support for machine learning acceleration, so that more AI applications can be implemented at the industrial edge.
2 High flexibility: The processor needs to provide rich interface resources to provide various wired/wireless interconnections and excellent scalability for different applications.
3 High reliability: This is a “must-have” for industrial application development, and the support for reliability cannot be ambiguous.
4 Complete development ecology: In addition to the chip, whether the supporting software development tools are available is increasingly becoming the “soft power” that developers should focus on when choosing a hardware platform.
02 NXP’s solution
If measured according to the above “core” standards, NXP’s i.MX 8M Plus application processor is obviously Mr. Right in the minds of developers.
i.MX 8M Plus System Block Diagram
First of all, in terms of computing power, the i.MX 8M Plus is based on a quad-core Arm Cortex-A53 with a main frequency of up to 1.8GHz, and integrates a neural network acceleration unit (NPU) to provide a computing power of up to 2.3TOPS. As the first i.MX processor with an integrated machine learning accelerator, the i.MX 8M Plus provides powerful performance support for machine learning at the edge – with the integrated NPU, the i.MX 8M Plus can simultaneously detect multiple complex neural network tasks, including human action and emotion recognition, multi-object monitoring, and recognition of over 40,000 words. This has obvious advantages in running machine learning tasks on edge devices, reducing dependence on cloud computing resources, protecting user privacy, and providing differentiated user experience.
Secondly, the high-speed interface and storage interface resource configuration of i.MX 8M Plus is also very “luxury”, including: 2 x USB 2.0/3.0 Type C, PCIe Gen3, 3 x SD/SDIO 3.01, 2 x Gbit Ethernet with EEE, AVB, IEEE 1588, one of which supports TSN for high-precision and low-latency control… In addition, 2x CAN/CAN FD interfaces are integrated. At the same time, i.MX 8M Plus also supports interconnection with various wireless communication devices such as 5G, WiFi, and BLE.
In addition, in terms of reliability, the i.MX 8M Plus has carried out special optimizations on key nodes of the system, such as the DDR interface supports error correction code (ECC), and supports the requirements of the SIL standard at the system level. Moreover, the 14nm FinFET manufacturing process provides i.MX 8M Plus with high performance and low power consumption, while also bringing a lower soft error rate.
Finally, in terms of software support, NXP can provide support for Android, Linux and FreeRTOS operating systems. At the same time, NXP’s eIQ™ software platform provides a series of libraries and design tools related to machine learning applications, aiming to help developers quickly deploy edge machine learning functions based on i.MX application processors and other NXP MCUs. At the same time, the eIQ tool is based on open source technology and is fully integrated in NXP’s Yocto development environment, helping users to better carry out system-level design. In addition, iMX 8M Plus can also support Real-time Edge industrial open source SDK to rapidly develop real-time applications; NXP has also ported OpenWRT to i.MX 8M Plus to achieve perfect support for industrial gateway products.
03 A new generation of industrial edge computing gateway
The strong strength of i.MX 8M Plus in industrial edge computing has been gradually revealed and released in the market.
Not long ago, Lubangtong released its new generation of industrial-grade edge computing gateway series products, among which the EG5120 based on i.MX 8M Plus is particularly eye-catching, which has the function of global 5G connection (supporting 3GPP 5G version 16 standard), It has 2.3TOPS computing power, supports AI applications, and can be widely used in industrial IoT application scenarios such as smart cities, smart manufacturing, and smart energy.
A variety of industrial-grade edge computing gateways released by Lubangtong (Image source: Lubangtong)
Specifically, the functions of the EG5120 include:
• Highly stable 5G / 4G / 3G cellular connection covering global frequency bands
• High-performance engine for running complex applications
• Meet the flexible high-performance interface requirements of industrial sites, including 2 Gigabit Ethernet ports, 2 RS232/RS485 (software configurable), 2 DI, 2 DO
• Dual SIM card redundancy backup, built-in 5G/4G wireless communication module
• Industrial grade design, high anti-interference performance, suitable for high temperature and wide working environment (-40 ~ 70°C), very suitable for industrial applications
• Equipped with RobustOS Pro IoT gateway operating system, support based on Docker container technology, support security features such as secure boot, system upgrade digital signature authentication
• Over 50,000 apps available in the Debian ecosystem
• Equipped with Neural Processing Unit (NPU), the computing power reaches 2.3TOPS, and supports machine learning
• Supports VPN options such as IPsec, OpenVPN, GRE, L2TP, PPTP, DMVPN, etc.
• Support RCMS (Lubangtong Device Management Cloud Platform)
It can be seen that behind the many powerful functions of EG5120, there are i.MX 8M Plus, especially in edge machine learning, dual Gigabit Ethernet support, video encoding (including H.265) and decoding, 3D/2D Graphics acceleration and audio and voice processing, providing hardware security mechanisms, etc., provide unique value.
04 Fully empowering industrial edge intelligence
In a word, the development of the Industrial Internet of Things has spawned industrial edge computing gateways, and it has also driven the demand for related industrial edge intelligent application processors. The iMX 8M Plus that came into being is the choice for this trend.
It is worth mentioning that for the personalized design requirements of different application scenarios of industrial edge computing gateways, NXP’s rich i.MX product portfolio can fully empower it. For example, another EG5100 industrial edge gateway released by Lubangtong in the same period uses NXP’s “field-proven” i.MX 6ULL high-efficiency and cost-effective application processor, providing an industrial solution that supports global 4G connections. IoT solutions.
Obviously, the technological empowerment journey of i.MX application processors in the era of industrial edge intelligence has begun, and it will accelerate forward.