Software-Defined Development Shift-Left: Start Development Earlier in the Age of Software-Defined Products

Related Vendors

Siemens Digital Industries Software GmbH

FAULHABER Drive Systems

Wuxi InfiMotion Technology Co., Ltd.

Software-defined products are currently changing the industry and calling many an established process into question. New dimensions of complexity are a challenge arising from this, and the digital twin is the answer.

The introduction of software-defined development approaches is triggering change in many industries, creating new scope for innovation. This can be seen very clearly in modern cars. Automatic software updates, which are regularly installed remotely in vehicles, can improve their performance, enable new functions or even rectify faults without the need for a workshop visit. The flexibility of a software-defined vehicle platform means that a manufacturer can save its customers a tedious task while offering new functions and improvements, thereby setting itself apart from the competition.

The transition to software-defined development has a major impact on a product program and affects all disciplines in the areas of design and construction. In my role at Siemens, I often have the opportunity to work with dedicated and experienced professionals whose technical expertise spans the entire spectrum we need to develop the smart products of tomorrow. This exchange constantly brings new technical, economic and other insights, especially with regard to the constant evolution of the digital world.

I recently spoke with Mike Ellow, CEO of Siemens EDA. It was about how much the digital landscape is changing and how software now determines the function of products in many industries such as automotive, aerospace and consumer electronics. In this post, I'd like to share with you a few key takeaways from that conversation; specifically, how software-defined development is changing companies' product and system development processes and how the comprehensive digital twin supports these new development approaches - specifically through the Digital Thread, which is the digital thread that provides the link between Electronic Design Automation, which is the automation of the design of electronic systems, the software and the rest of the system.

The Software-Defined Approach Increases Complexity

The transition to software-defined products leads to new complexity, as changing the functions of software can have cascading effects across several interdependent domains. For example, if the behavior of the on-board software of a future electric vehicle changes, this can affect the power consumption of the computing platform and thus the overall range. If such a software change significantly reduces the range of a vehicle, the battery capacity may have to be recalculated or even the installation space in the vehicle adjusted.

The increased focus on software development and differentiation is also impacting the organization of development programs, particularly for semiconductor devices, which form the physical basis for software functions and features. While the mechanical and electrical systems of products are still important, the focus of attention and investment is now shifting towards software and semiconductors.

Traditionally, vehicle manufacturers have used off-the-shelf circuits that are sufficient for their purposes and then developed the software according to the capabilities and limitations of these general-purpose hardware platforms. However, as software systems become more sophisticated and important, more and more companies are looking for customized semiconductor solutions that are tailored to specific software tasks and for development processes that support the co-design of hardware and software.

Until now, however, such customer-specific chips were out of the question for most companies: long lead times, high costs and the risks of developing application-specific circuits were arguments against it. Semiconductor manufacturers are now shortening their development times by using important innovations such as heterogeneous integration. In doing so, they combine inexpensive standard elements, such as memory, with smaller, less expensive and also less risky customer-specific circuits, thus optimizing the cost-performance ratio.

The Digital Twin Simplifies Holistic Software-Driven Development

Today, manufacturers of complex systems must consider the development and integration of different subsystems. These include

• Software applications and basic functions;
• Semiconductor components and assemblies;
• electrical and electronic systems including data networks;
• mechanical components and structures.

All of these elements require different technical skills and solutions, but should ultimately be brought together to create a uniform, attractive and user-friendly product from a single source. Manufacturers must master this task despite increasingly tight budgets and time constraints if they want to hold their own in increasingly competitive markets.

The pressure requires networked, agile and holistic development methods that facilitate the flow of information between interdependent teams; it requires investment in digitalization through the creation of a comprehensive digital twin. The comprehensive digital twin of a complex product or system connects all product models and data in the form of digital threads that enable the flow of information between development teams and project managers in all directions. The result is constant coordination between the work of the developers and the requirements of the designers.

Especially for products such as electric vehicles or aerospace systems, which are becoming increasingly complex, digital twins offer developers the advantage of a unified platform for visualizing dependencies between software, electronics and mechanical elements. Real-time synchronization helps to avoid costly delays by ensuring that a product meets the ever-changing requirements and legal regulations in every respect.

Beyond the work of individual teams, digitalization also enables better cross-functional collaboration by making data easily accessible to all stakeholders in a transparent form. This allows supply chain managers, software specialists and design engineers to share information in a common digital ecosystem, streamlining validation procedures and preventing unforeseen complications arising from changes. The digital twin also brings benefits for production preparation, as it enables early production planning, modeling and validation. This holistic approach reduces development risks and increases efficiency: high-quality, innovative products are on the market faster.

Better Software Development Through the Virtualization of Semiconductors

For software-defined products or systems, virtualizing the development and validation of semiconductor elements is essential. Previously, software developers waited for physical prototypes of the hardware before testing their code - this made the development cycle slower and increased the project risk. Today, digitalization allows the virtualization of semiconductor architectures and thus the creation of program code independently of the development and selection of hardware. This means that active software development can begin long before physical semiconductor components are manufactured.

Virtualization makes it possible to take dependencies between software and hardware into account when designing products before full integration has taken place. The accuracy of a semiconductor model increases steadily over the course of development, so software teams can continuously update their applications with the latest calculated semiconductor performance data, reducing the risk of integration errors. Virtualization also enables manufacturers to fully exploit the potential of heterogeneously integrated semiconductors by facilitating the optimization of processor cores, memory configurations and interfaces to meet growing computing power requirements at lower cost.

A networked digital ecosystem also ensures that the comprehensive digital twin remains up to date when properties or technical data of software or semiconductors change. As the software and semiconductor configurations evolve and mature, the electrical and mechanical teams can use the latest data for their design work and consider any potential impact of a software change on their systems.

To return to the example of an electric vehicle: The digital twin makes the consequences of a software change completely transparent throughout the entire development environment of a product. So when software developers work on the code to achieve a certain system behavior, they can optimize their changes to minimize the impact on other assemblies. Once the changes to the software have been completed, new requirements can be forwarded to the affected teams so that they can adapt their designs if necessary.

Using Complexity as the Key to Software-Driven Development

The growing influence of software-defined products is changing the industry and creating new opportunities for innovation and increased efficiency. However, the increasing reliance on software is also leading to new forms of complexity that require a more integrated approach to development. The convergence of software and semiconductor development has led to an increase in demand for customized chip solutions where the computing power is tailored to specific workloads. Digitalization, particularly in the form of the digital twin, has become a key tool for managing interdependencies. It enables working groups to collaborate and validate across departments in real time.

In the age of software-defined products, a company's success ultimately depends on its willingness to invest in digital infrastructure and the development of a comprehensive digital twin. The digital twin is the key to fast, lean innovation processes, reduced risk and better product differentiation. The industry is changing and evolving - those companies that leverage the new tech capabilities most effectively will define the next generation of smart, connected products.

*Dale Tutt is Vice President of Global Industry Strategy at Siemens Digital Industries Software.