Platform-Based Design and Test in a Programmable World
We live in a platform-based programmable world, according to Mike Santori, National Instruments business and technology fellow. Speaking at the NIDays Boston event last November, Santori described how our world has transformed into a huge programmable system, to which many terms have been applied: the Internet of Things, for example, or, in Germany, Industrie 4.0. It’s a world populated by smart systems, smart factories, M2M capabilities, and ubiquitous sensors, all accompanied by Big Analog Data (a term NI has trademarked), and all of which might fit under the label “cyber-physical systems” (CPS).
NI Business and Technology Fellow
Santori quoted Vijay Kumar, assistant director for robotics and CPS, Office of Science and Technology Policy at the White House, as saying “CPS will transform entire industrial sectors, including transportation, healthcare, energy, manufacturing, and agriculture. There is an urgent need to develop design methodologies that will provide real-time, guaranteed performance.” CPS, Santori said, represent areas of intense research at NI.
Santori described a world in which platforms like Android and iOS mobile devices have taken on functions once handled by dedicated products such as the wristwatch and GPS. Special-purpose devices are going away, Santori said, being replaced by virtual implementations. Windows and Intel-based hardware offered a standard platform in the personal computer space, and HTTP/HTML provided a standard software platform for the World Wide Web, he said, and a similar evolution has taken place in measurement and control. The NI LabVIEW RIO architecture—comprising a processor, FPGA, and analog and digital I/O—is the platform on which virtual instrumentation is built, he said.
In his NIDays Boston keynote address, Santori reinforced points that had been made throughout NIWeek in Austin last summer. In an interview following his Boston presentation, Santori said that since NIWeek, “Probably the topic that ended up generating the most new interest is Big Analog Data—that struck a chord with a lot of people. A number of customers have come to talk to us about building large, distributed systems where they are generating, collecting, and trying to manage a lot of information.”
Duke Energy is one such customer, he said. The company has deployed Compact RIO systems to analyze data in a machine-condition monitoring application (as described at NIWeek). Santori said that in the mass media, big data is considered an IT topic, whereas “in our world it’s really more about machine-operating data, measurement data, and crash test data—automotive engineers collect a lot of information.”
In fact, he said, the challenge of Big Analog Data extends beyond the data itself. “It’s actually about much more than the data—it’s about being able to manage all the devices that are collecting the data, processing the data, and presenting the data. So when we look at the whole Big Analog Data architecture, we see lots of opportunities to deploy analysis and data reduction and data processing out into intelligent nodes.”
Santori then returned to the topic of CPS. “There’s a set of connected trends that almost all lead back to the Big Analog Data trend—lots of data and lots of connected distributed systems that have a lot of intelligence in them. That’s why we like to use the term cyber-physical system”—a higher level term that basically envisions not just a bunch of connected systems but rather a system of systems. And in addition to data collection, processing, and management, designers of such systems need to take into account such issues as security and reliability while keeping pace with evolving bus and network standards and mobile technologies.
Concerns about reliability and security are affecting where companies store the data they are collecting. “By and large, most companies—by nature of the data that they are collecting—will tend to keep it on their own servers within their own firewalls,” Santori said. “They don’t broadly accept public data storage yet because they are concerned about security.” Issues remain, he said, with regard to a lot of data remaining on local machines—especially isolated test systems or isolated data-acquisition systems. But increasingly, he said, “as everything becomes networked through either wired Ethernet or Wi-Fi, the data is ending up on more centralized servers, but typically within the company’s own operations.”
In describing how NI has met the challenges leading up to today’s programmable world, Santori said, “We’ve followed the lead of Moore’s law since the days of the PC. What really enabled NI was basically being on top of the adoption of the PC and leveraging off-the-shelf technology.”