“What are the characteristics of software and hardware development in the Internet of Things era? The first is that costs are more sensitive. The IoT market has huge potential, but the Internet of Everything means the diversification and fragmentation of application scenarios, which leads to relatively limited application scenarios for each chip, and the mass production of chips is severely fragmented. This makes the cost of chip customization an important factor affecting its feasibility and popularity.
Chip shipments based on the ArmCortex-M architecture have so far exceeded 70 billion and are still growing. According to MordorIntelligence, the average compound annual growth rate (CAGR) of IoT chips will reach nearly 15% in 2026. To capitalize on this growth opportunity, Arm is focused on ensuring that the ecosystem remains at the forefront of accelerating IoT innovation.
If Huawei Hongmeng system and Google’s FuchsiaOS represent the operating system from the mobile phone era to the IOT era. Then ArmTotal Solutions for IoT recently launched, including ArmCorstone, Arm virtual hardware target, and ProjectCentauri three key parts, which have revolutionized embedded development and brought the software and software in the Internet of Things era. Hardware development method.
What are the challenges facing software and hardware development in the Internet of Things era?
What are the characteristics of software and hardware development in the Internet of Things era? The first is that costs are more sensitive. The IoT market has huge potential, but the Internet of Everything means the diversification and fragmentation of application scenarios, which leads to relatively limited application scenarios for each chip, and the mass production of chips is severely fragmented. This makes the cost of chip customization an important factor affecting its feasibility and popularity. The severe fragmentation of the Internet of Things has led to extremely high cost sensitivity, including the cost sensitivity of chip tape-out and the cost sensitivity of the application market. In addition, the rapid development iteration and launch of chips have also become the focus of customers. Customers want to bring products to market quickly, but are generally faced with cost pressures and a shortage of IoT talent. There are not enough IoT software developers in the market, even if there are a few developers, they have to face many areas that they are not good at, such as security issues in the IoT field, connectivity issues, machine learning capabilities, embedded design, application design and development for cloud services.
The second is customizability. The diversification and differentiation of IoT application scenarios have led to the need for customization of chip products. The diversification and differentiation of IoT application scenarios make different application fields have different requirements for the details of chip solutions. Due to the extremely fragmented downstream application fields of the Internet of Things, the chip design is also more diverse, and the demand for customization is strong.
The IoT market has also spawned more design demands, resulting in lower market entry barriers and a substantial increase in the number of chip design companies. Many atypical customers (such as Internet companies) that are different from traditional chip manufacturers in the past have begun to emerge. This type of of customers often lack sufficient embedded development experience.
Mohamed Awad, vice president of the IoT and Embedded Group at Arm, reviewed the evolution of IoT and divided it into three stages. He believes that the Internet of Things is entering the era of intelligence. Although the intelligence of the Internet of Things has great potential, from an industrial point of view, the speed of development is still not fast enough. So, what is holding back the development of the Internet of Things?
MohamedAwad believes that there are three main reasons:
1. Slow product design, due to the continuum of development, it takes five years or more to design the latest technology into a product.
2. Inefficient software development, IoT development needs to be developed and tested based on physical hardware, and embedded development follows the same way.
3. Lack of scale, which prevents software and services from being easily used across platforms.
Today, there are millions of mobile apps and thousands of mobile phone models on the market. If every mobile app developer needs to buy all the mobile phones to adapt the app, the number of app developers will be greatly reduced. The same is true in the field of the Internet of Things. If the software for the development of the Internet of Things also needs to adapt to various hardware terminals, the development of the Internet of Things will become difficult. In fact, many atypical developers, although they are not embedded developers, are also innovating the Internet of Things, and they urgently need Arm to solve the pain points or challenges they face in these development.
This is why Arm has released a comprehensive solution for the Internet of Things (ArmTotal Solutions for IoT), and only Arm has the potential to solve these problems. There are several reasons: First, there are currently more than 70 billion Cortex-M-based chips shipped to the market, and it can be said that most IoT terminals run on the Arm architecture; second, Arm has been committed to accelerating product design. hardware and software development, and has introduced continuous integration and scalable cloud-based testing; third, Arm has experience working with ecosystem partners to address these challenges.
virtual hardware target
According to reports, the Arm IoT comprehensive solution is a complete set of solutions designed for specific use cases, allowing developers to focus on the really important part, which is innovation and differentiation across different applications and devices. It has everything you need to simplify the design process and product development, including hardware IP, software, machine learning models, advanced tools such as new virtual hardware targets, application-specific reference code, and sources from the world’s largest IoT ecosystem support.
Arm Total Solutions for IoT will take a more modern approach to software development and easily innovate for the diverse hardware that builds the Internet of Things, said Mohamed Awad, vice president of Arm’s IoT and Embedded Business Unit. He particularly emphasized that Arm will completely change the way of IoT software development from the bottom, and at the same time will change the IoT market. Achieving this will not require them to transition to embedded developers. Arm IoT comprehensive solutions will accelerate product design cycles by up to two years.
The Arm® Total Solutions for IoT (ArmTotal Solutions for IoT) consists of several components, one of which is the Arm Virtual Hardware Target.
Arm virtual hardware is an industry first that enables software development to be done in the cloud, revolutionizing the way IoT software is developed. The new cloud-based service provides a virtual model of the Corstone subsystem, allowing software development to take place without a physical chip. Arm Virtual Hardware brings modern and agile software development methods to IoT and embedded platforms, including continuous integration/continuous development (CI/CD), DevOps and MLOps, without investing in complex hardware farms. It also means that millions of software developers don’t need to transition to embedded developers to innovate for IoT development. .
MohamedAwad also said that the concept of Arm virtual hardware is to combine the existing simplified mobile application development model with a cloud-based form so that software developers and developers can better develop IoT applications. The power of a set of solutions is something that FPGAs cannot do. FPGA solutions are typically used for entry-level embedded development, and developing a dedicated device for testing at scale is difficult and expensive. In contrast, the Arm virtual hardware target is based on the cloud and is provided to third-party software manufacturers or software developers for use. They may have been developing smartphone applications or cloud service applications in the past. Arm virtual hardware can help them expand. Application development in the field of Internet of Things.
Through accurate models of Arm-based SoCs that provide mechanisms for simulating memory and peripherals, software development and testing can now begin before chips are complete. As a result, typical product design cycles can range from an average of five years to a maximum of shortened to three years. With these new tools, machine learning developers no longer need to transform into embedded developers, they can use virtual hardware to optimize modules in the cloud. They can run continuous integration workflows and validate their algorithms on virtual hardware.
This allows Arm chip partners to obtain customer feedback on the chip before the chip is taped out, while assisting the entire IoT value chain, and can easily develop and test code based on the latest IP before the chip is released. On the left, there is a machine learning development workflow where neural network optimization for the edge runs in the cloud, on virtual hardware. To the right, in the middle section is the software development integration. The actual development will take place on virtual hardware, so you don’t need to worry about your fleet or hardware farm. Until you’re ready to deploy to real hardware, then you’re ready for full deployment. This level of optimization across workflows not only increases productivity, but also allows more new players to enter the market, accelerating innovation.
MohamedAwad also said that from a software perspective, most of the focus and initial use cases of Arm virtual hardware are focused on DevOps, which is continuous integration/continuous development (CI/CD) for software. And machine learning DevOps is to optimize machine learning models and deploy them on terminal devices. In today’s market, in addition to the Internet of Things, machine learning development in other fields is basically based on the cloud, and the development of machine learning requires special expertise. When the Arm IoT comprehensive solution introduces the cloud-based development approach to the IoT, it can be expected to see more and more intelligent applications on the IoT. He took Amazon as an example, and Amazon applied Arm virtual hardware to test the Alexa wake word. The Alexa wake-up test can be used on 150 different devices more quickly than Amazon has done in the past. Another example is Himax, who are using Arm’s comprehensive IoT solution to accelerate their new AI development schedule, and their dedicated Arm virtual hardware for the new processor is available to developers before the chip is complete.
It all starts with Arm Corstone, which is a pre-engineered, pre-integrated, and pre-verified hardware subsystem. Corstone is built around specific use cases, and Arm provides the technology directly to chip design partners. Corstone is designed specifically for chip design partners and OEMs to focus on differentiation. This is done by combining the CPU, NPU, system IP, and other key technologies into a pre-integrated, pre-validated, and ready-to-run subsystem. It is understood that Corstone has accelerated the time-to-market process for more than 150 design projects for Arm chip partners. Arm virtual hardware is available now on AWS Marketplace, with a planned launch in China in 2022.
Cortex-M’s Ecosystem Initiative Project Centauri
Another part of ArmTotal Solutions for IoT is ProjectCentauri, an ecosystem initiative for Cortex-M. The project aims to achieve a similar contribution to the Cortex-A ecosystem by Project Cassini by providing a set of device and platform standards and reference implementations for device boot, security and cloud integration for the broad ArmCortex-M software ecosystem. ProjectCentauri’s API includes support for PSA certification and Open-CMSIS-CDI, a set of standard cloud-to-device specifications that minimize the development effort required to launch different cloud solutions and real-time operating systems. ProjectCentauri will reduce engineering development costs, accelerate time-to-market, enable large-scale IoT deployments, and strengthen the security of the Cortex-M ecosystem.
This is the product roadmap for Arm’s IoT comprehensive solution, the first of which has been launched. It is aimed at machine learning use cases including keyword recognition and is based on Corstone-300 using Cortex-M55 and Ethos-U55.
With Arm IoT Total Solutions, Arm’s vision is to accelerate product development and improve ROI for everyone, including OEMs, IoT product and service providers, and a new generation of IoT that combines traditional and modern IoT development practices Developer. MohamedAwad said that the traditional development process, including IP selection, chip design, and then hardware manufacturing, while software development can only be started after driver development, board support package and application development, and the whole process can take up to five years. . After empowering developers with the Arm IoT comprehensive solution, software development can be started at the same time as hardware development, and the product launch cycle can be reduced from five years to three years.
Arm’s smartphone business is based on Arm’s embedded technology, and the current IoT business is also derived from embedded technology. Arm has benefited a lot in the smartphone era and has accumulated a lot of experience. The take-off of Arm smartphones is based on the ecosystem being able to develop software in a common way, thereby driving innovation in hardware.
Now, this situation is happening again in the field of IoT. MohamedAwad believes that the most important point of Arm’s comprehensive IoT solution is to transplant the innovation driven by the mobile field to the IoT field at the same level. By reducing the investment in software and expanding the market that can be served, it will promote the investment of vast software resources to the Internet of Things. IoT market. MohamedAwad believes that the new Internet of Things economy will be driven, and its scale and prosperity will definitely surpass the smartphone market.
“By revolutionizing the way systems are designed, Arm is uniquely positioned to drive a new IoT economy that is in shape, speed and scale comparable to the smartphone app market economy,” said Mohamed Awad, vice president of Arm’s IoT and Embedded Group. Comparable. Arm IoT Total Solutions transforms the way we deliver key technologies across our ecosystem and demonstrates our significant and ongoing investment in software that empowers developers to innovate and expand global reach.”
The author believes that Arm is reconstructing its own products in the face of the particularity of software and hardware development in the Internet of Things era. Arm hopes to replicate its development methods in the mobile field to the Internet of Things field, helping more customers shorten product design cycles by up to two years, and better cope with the challenges brought by open source architectures such as RISC-V. Finally, the author believes that with this comprehensive solution as a design, Arm’s ecological partners can better focus on innovation and improve their own return on investment. Through standardization, partners can better differentiate themselves in the segmented field, and they can make full use of their investment to solve new challenges in the market without wasting on repeating the work that has already been done.