San Francisco. Erik Volkerink, Ph.D., chief technology officer of Flextronics, outlined the future of global technology to attendees of the Test Vision 2020 Workshop, held in conjunction with SEMICON West. Although he now focuses on products and systems instead of semiconductors, he has extensive experience in semiconductor test, having served as chief technology officer at Advantest and, before the Advantest acquisition, Verigy. He has also been an active participant in the Test Vision 2020 Workshop.
“It's good to be back here,” he said, noting that in his current position, he can offer outside perspective.
Flextronics, he said, is a $24 billion company with 50,000 employees, including 3,000 design engineers who design products for customers, who affix their own brands. Flextronics, he said, is involved in the full product lifecycle—not just manufacturing, but design, distribution, and repair as well.
He cited a list of product types handled by Flextronics: notebooks, servers, printers, smartphones, data cards, set-top boxes, medical drug-delivery devices, wearable consumer-health devices, kiosks, and camera modules. To handle such diverse devices with a relatively small design team, he said, it's important to look at trends and generalize requirements.
The future will be populated by smart devices and systems, such as the smart grid, he said. If you pick a random noun and put the word “smart” in front of it and Google the combination, he said, you'll find several startups already engaged in designing the smart widget. He cited as an example the “smart floor“—it can detect when an elderly person has fallen down, for example. Similarly, you'll find smart curtains, shutters, and lighting, which can be controlled remotely via a mobile app.
In addition, he said, smart automotive technologies will detect your smartphone,you’re your car can replicate the phone's interface. And within the next three or four years cars will come with 20 or more cameras.
In addition, he said, smart glasses will drive the next level of convergence. Smart glasses don’t need screens, he said, and will completely disrupt human-machine interfacing. Similarly, smart gloves will provide tactile feedback and support intuitive robotic surgery. And smart sensors will proliferate. “It's quite amazing, the number of companies in Silicon Valley designing smart body sensors to detect heartbeat or other medical issues.”
He cited additional smart devices and systems, ranging from social shirts and the connected home to smart shoes, smart vending machines, and smart identity biometrics, which can replace passwords with a heart-beat-based ID. He cited IDC figures predicting that smart products will account for $2 trillion in revenue in 2015.
The trend toward smart devices entails considerable disruption while presenting design and engineering challenges, he said. If you are a clothing manufacturer who wants to make smart clothes, he said, you'll need to hire your first engineer. “A lot of market segments are being disrupted right now,” he said, “presenting opportunities for semiconductor test companies.”
He noted that even as product complexity increases, time-to-market pressure goes up exponentially. Test, he said, is an important enabler in relieving the resulting stress. And chip-level test granularity can help add value to system-level test.
He cited system-level diagnostics, testing, and validation and emphasized the importance of testing across abstraction levels without false positives. He also commented on the challenges of working with heterogeneous technologies and nontraditional integrations of functionality.
He also advised, “To do system level test without false positives or escapes, you need to find the balance of tolerating defects. You need to tolerate things that are wrong—otherwise, yield suffers.”
He noted that an old manufacturing paradigm was capability-based, in which companies sell based on low cost; today, the paradigm is component-based, with companies selling time-to-market. Future product-design and engineering flows will be based on platform strategies. Such strategies will enable efficient product test, component test, platform test, and system test. He called for collaboration to support prequalification for testability, manufacturability, and diagnosability, adding new types of hardware/software co-testers will emerge.
The next generation of products, he concluded, will combine electronics, fluidics, biology, and mechanical assemblies, and industry needs to collaborate to take advantage of emerging opportunities.