Hot Technologies In Cold Weather

It is a busy week for verification and software development. DVCon in San Jose; Embedded World in Nuremberg, Germany; and Mobile World Congress (MWC) in Barcelona, Spain are all happening at the same time. I ended up covering embedded software in Germany (as I also had a paper on “Shift Left” here). At chilly minus 1° Celsius in the morning, the technologies had to be pretty hot to warm me up! My top five take-aways from Embedded World ‘19 were about:

• Security and safety: Very different in themselves, but both really hot topics that impact all of us
• Artificial Intelligence and Machine Learning (AIML): making me want to learn more myself
• Automotive: Making our soon-to-be self-driving cars the coolest gadgets we can own
• Ecosystems in processors: That may have subtly changed already
• The ability of counting being a necessary requirement in our business: Apparently, industrial revolutions have reached its fourth wave; computing its fifth, and Wi-Fi its sixth.

You have to read on to decipher that last one. Can you hear me now?

There has been a steady demise of embedded conferences in the US. In contrast, Embedded World in Europe seems to continue to thrive. It was as busy as ever, with six halls of related embedded activities. It includes items like displays, packaging and single board computers, so I could not relate to all of it really—but the areas for automotive, embedded development tools, microcontrollers and everything IoT looked extremely busy. The conference has added an SoC track this year, so there is a good mix of hardware and software developers here.

First, safety and security were the stand-out topics when entering the halls. Everybody was talking up horror stories about what could happen if our gadgets are not secure and safe, from being able to purchase hacks to failures in day-to-day used items that have to be avoided. One tool not being compliant to the right standards can seriously jeopardize hardware/software developments. There was a slew of certifications of compilers and other development tools to allow ISO26262 compliance. And software security was a big topic too—Arm announced certifications as the fourth component of their Platform Security Architecture (PSA) and vendors were showing code analysis during compile time as well as run-time security with operating systems like Green Hills’ INTEGRITY showing their containers with different levels of security. This allows critical areas like the life-preserving functions in medical applications and the safety-critical applications in a car to run on the same OS; think your camera and lidar continuing to work safely and securely in their containers still while navigation and entertainment can crash without bringing them down. Cadence recently announced a strategic partnership with Green Hills; I am looking forward to that bearing fruit.

Second, AI/ML was present everywhere—in the cloud, at the edge, spanning from severs through networks and microcontrollers. For instance, Cadence has just announced our support of the verification suite for the new Arm Neoverse N1 Platform to advance the cloud to infrastructure market. We were showing Tensilica-based visions applications at our booth, and the conference keynote by Mathworks’ fellow Jim Tung “Developing Game-Changing Embedded Intelligence” offered interesting insights into the development of AI/ML algorithms.

Third, the clustering of ecosystems around processor architectures was very visible, both directly with the presence of Arm and its licensees, Intel and a RISC-V ecosystem booth—but perhaps even more interesting, they were present in a subtler way with tool vendors like Lauterbach and IAR simply adding demos for RISC-V tool support at their booth. I had written about the “Game of Ecosystems” before and more recently about Arm Techcon, and the jury is still out here. In my discussions with colleagues and market experts, the big question was whether value-added tool support for RISC-V will find users willing to pay for them, given that one big draw of using RISC-V is to start from an open source base. But then again, the hardware implementations of architecture variations need to be verified, and software needs to be developed securely and safe. It’s a fascinating space to watch.

Fourth, Germans love their cars and automotive was a huge central element again. ADAS and driver assistance are huge market drivers. We at Cadence showed results of our partnership with Dream Chip at our booth, highlighting Tensilica IP. And in my discussions on the tool side, safety certification of development tools was not a nice-to-have but a requirement. And users are working out ways how to apply our verification tools to reach the different ASIL certifications of safety.

My fifth takeaway, somewhat less technical, is about the different numbering schemes. All the different application domains count somewhat differently. The Huawei Wi-Fi 6 advertisements could not be missed when arriving at Frankfurt airport in the customs hall, probably counting on an audience switching there to Barcelona and coming to Frankfurt. Of course, cellular networks are about to upgrade from 4G LTE to 5G. Then Arm was welcoming people to the fifth wave of computing—that is, to quote Arm’s CEO Simon Segars, “an era of computing that will be data-driven, with traditional algorithmic computing giving way to data flowing through machines and decisions made based on what data is telling us.” It is following the four waves of mainframe computing, personal computing and software, the Internet and mobile and cloud computing. (Shout-out to my friends from Scotland and Minnesota at the Arm booth with whom I figured this out without googling it.)

In contrast, the industrial domain has only counted to four revolutions so far. It also goes back the furthest, starting from mostly agrarian, rural societies becoming industrial and urban in the 18^th and 19^th centuries, continuing to the period of growth until 1914 marked by the expansion of new industries such as steel, oil and electricity, using electric power to create mass production. The third revolution started with the adoption of computers and automation in the 1980s and is enhanced now to Industry 4.0 with smart and autonomous systems fueled by data and machine learning. (It’s probably time to admit at this point that I am the son of a history teacher…)

No matter whether your number is 4, 5 or 6 though, its undoubtedly exciting times in electronics. Now can we move Embedded World to the summer please? Nuremberg is nice, but I am freezing.