What does the industry represent for the semiconductor sector? As they argue in Micron, high-performance, energy-efficient memories and storage are now a natural choice to meet the need for all-round connectivity and smart services and products.
"Today's factories are complex, have machines that are connected 24/7, requiring statistics, remote monitoring, automation and autonomy and rapid reconfiguration with little or no downtime," explains the management of the US giant which, with industrial application providers such as Advantech and Kontron, has launched theIndustrial Quotient (IQ) Partner Program for the sector and has invested billions of dollars in wafer fabrication to extend product life beyond a standard life cycle, something that is valuable to manufacturing customers. We asked Micron's management a few questions about the relationship between electronics and industry.
How important is the manufacturing (machinery and automation) market for Micron and who are your stakeholders in this sector?
Industry is an important market for Micron and we have been proudly supporting this sector for decades. During this time, we have seen a significant evolution of our manufacturing and industrial customers. Connectivity everywhere, smart services and products, and challenging environmental requirements have made our high-performance, energy-efficient memory and storage devices a natural choice for industrial customers looking for better results by moving computing closer to where data is created. Highlighting how important this market is to us, in 2020 we launched our Industrial Quotient (IQ) Partner Program that provides industrial customers with robust memory and storage for applications in transportation, factory automation and more. The program includes companies that provide a wide range of industrial applications such as Advantech, ATP Electronics, Greenliant, Innodisk, Kontron, Mercury Systems, Viking Technology and SMART Modular Technologies. We view our industrial quotient as a set of guiding principles that ensure our industrial-grade solutions not only meet customers' functional requirements, but are also reinforced to withstand extreme conditions and ensure long-term quality and reliability, all while keeping product lifecycle management simple. This requires investment in both process and product technologies to achieve this balance required by the automation industry and Micron has put together a portfolio of high quality storage and memory devices to meet this challenge. Micron has also invested billions of dollars in wafer fabrication to improve product longevity with the ability to extend product life well beyond a standard life cycle. This is particularly valuable for customers such as those in manufacturing who require memory and storage solutions with the longevity to meet the requirements of their industrial-grade equipment, which tend to have very long life cycles.
Today, when we talk about electronics for the manufacturing industry (machinery and automation) we talk about intelligent production...
Today's manufacturers are benefiting from the broader trend of extending artificial intelligence (AI) capabilities into new locations, such as a robot on the factory floor, on top of the foundation of ubiquitous 5G connectivity and high-performance memory, storage and computing hardware. We refer to this approach to computing as 'the intelligent edge'. It is helping us to realise the potential of what has been called Industry 4.0, which Deloitte eloquently explains as "a new industrial revolution, combining advanced manufacturing techniques with the Internet of Things to create production systems that are not only interconnected, but communicate, analyse and use information to drive further 'intelligent actions' in the physical world" . Today's factories are complex, 24/7 connected machines that require statistics, remote monitoring, automation and autonomy and rapid reconfiguration with little or no downtime. This means that electronics in machines and automation must be robust, of high quality, able to withstand high temperatures and have long life cycles. These requirements help to keep the factory running and simplify downsizing operations. All types of electronics are required, including sensors, connectivity devices (wired and wireless), microcontrollers, connectors and FPGAs (Field Programmable Gate Arrays) ranging from custom I/O controllers to sophisticated accelerators. In addition, central control systems and computer numerical control (CNC) require higher performance processors to drive decisions and actions in factories that predict failures, detect slowdowns or empty lines, and program machines remotely. Advances in memory and storage density, performance, power and cost enable these systems to efficiently store, analyse and communicate decisions. Ultimately, these actions can help increase productivity and reduce accidents on the factory floor.
Which electronic solutions have contributed more than others to making factories smart in recent years? How?
The evolution of manufacturing depends on a wide range of different electronic devices. But when we start talking about smart factories, we are actually talking about artificial intelligence that requires immediate and proximal access to memory, storage and extended capacity computing hardware. Therefore, the large number of low-cost, high-performance processors and high-end standard processors (FPGAs, AI CPUs, IPC processors) has contributed enormously to making factories smarter. The increased availability of a high-performance processing infrastructure tailored to often difficult usage environments has enabled new approaches to data analysis, machine learning and artificial intelligence. But, of course, these new differentiators cannot occur unless computation is matched with the corresponding memory and storage infrastructure. Without advances in these memory and storage components, it would be difficult to keep up with the DRAM performance and density required for these smart factory applications. Micron's continued investment in supporting the manufacturing industry has made us a trusted advisor for industrial memory and storage architectures, as well as the expertise to deliver solutions that meet or exceed our customers' application needs and the reliability and endurance requirements of industrial use cases. Our initial history in the IPC space has expanded with chipset partners such as NXP, Intel, Xilinx, Renesas, Texas Instruments, ST Microelectronic and Lattice to name a few. Our technical efforts with chipset partners help ensure that our memory and storage solutions work with processors, chipsets, FPGAs and the like. Together, these electronic components work to form a solid hardware foundation that makes factories smart by powering intelligent algorithms and sensors that automate and simplify manufacturing.
Looking at electronics for industrial robotics, what are the most innovative technologies in this field today?
As the field of robotics continues to evolve, we are seeing a growing and constant demand for computing infrastructure that supports sensors, software intelligence, analysis and decision-making, for which artificial intelligence capabilities are critical. The data inputs to robots are immense, flowing from online sources, local sensors, voice commands, tactile and traditional key inputs, and the robot must respond to these inputs quickly and accurately. This is where the electronics community has stepped in to meet these requirements. Processors have had to reduce costs, but increase speed, power and performance to keep up with the demands of new AI technologies. Input and output (I/O) functionality was either integrated into the processor or developed separately via FPGAs. In addition, customised application-specific integrated circuits (ASICs) have grown in complexity as the number of stimuli and data operations increases exponentially. As processors, FPGAs and custom ASICs evolve, memory and storage solutions must advance in parallel to keep up with the system-level demands that AI applications place on the processing infrastructure. Memory performance must be fast enough to react to sensors and make decisions to avoid problems that could cost the business money. In addition, storage must be energy efficient and fast to avoid the bottleneck of critical data and decisions. Finally, every aspect of the computing infrastructure must be robust enough for demanding usage environments such as the factory floor. At Micron, we have continuously innovated our memory and storage to reduce power, lower cost per gigabyte, increase performance and ensure our solutions are resilient to extreme industrial conditions. We have also increased the capacity of our DRAM, which helps the robotics industry keep pace with the size of software code and large data stores for localised analysis and decision-making, processing data where it is created at the intelligence edge. We continue to strive to meet the rigorous memory and storage requirements as they evolve in the future of automation and robotics.
You might also be interested in:
Micron: the focus is on memories and storage for Data Center
