The Democratization Of System Design…

…And the fourth major era of computing.


One of my favorite technology columnists, Andy Kessler, argues in a new Wall Street Journal column that we’re entering a fourth major era of computing:

“The original mainframe computer of the 1960s automated back offices and transactions, bringing efficiency and lowering costs. That cycle ended in the early 1990s when the personal computer picked up steam. By the mid-1990s the Web was helping to automate interaction between companies and their customers, allowing people to buy goods and services through the magical Internet.”

We have just entered the mobile era because there are now more users of mobile devices than there are of desktop devices.

“Having a personal computer in your pocket is changing the tech world because, unlike a computer, a smartphone is always there when you need it,” Kessler writes.

He sees companies working diligently to perfect the mobile-cloud infrastructure and once that happens, technology spending will accelerate and productivity and profitability will rise.

A different kind of cycle
He argues that for a couple of decades now technology cycles have replaced the old four-year economic cycles that businesses set their watches to. Technology has smoothed those cycles in which inventory builds up to oversupply, prices fall and companies pull back and retrench for a time before the cycle kicks in.

Electronics companies have been enabling this new era of computing for a few years now, but challenges remain. They remain because one of the key structures that will drive this new era is the Internet of Things (IoT), a brave new world of system design that we’re just getting our heads (and engineering teams) around.

Internet of Things applications require a new cycle of innovation for engineering teams. Over the decades, teams have grappled with these cycles within cycles. They’re gone through engineering cycles driven by performance and cost; cycles driven by integration and costs; Makimoto’s Wave cycles of customization swinging to programmability and back again.

The electronics innovation cycle, if you will, that we’re entering is perhaps the most complex engineers have ever encountered. They’re pressured to deliver devices quickly that optimize power, performance and area and, for IoT, hit historically low cost points.

Looking back, it’s a predictable path, traced more or less to the invention of the transistor. But this cycle — following the achievements of the past 60 to 70 years — brings with it something new: the democratization of technology innovation.

Disruptive influences
Open-source software and hardware, integrated and widely available SoCs and programming languages, have sparked an era we can only glimpse at this point.

Think about the rise of the drones. This technology is enabled by what I just mentioned—open, available and approachable technologies in the hands of clever people. Author, entrepreneur and former Wired magazine editor Chris Anderson, in a keynote address last week at ARM TechCon, noted that the drone phenomenon has occurred well outside the purview of big aerospace companies like Boeing.

In describing his own drone company, Anderson said:

“It’s two years worth of actual company building based on an open source community that’s creating the PC to the aerospace industry’s mainframe. We’re reinventing how we think about putting things in the air. We look more like a smart phone than an airplane.”

This is the democratization of electronics innovation.

In the process, engineers in the trenches will accelerate the cycle by thinking more about purpose and power and thinking less about device or system type.

My colleague Chris Rowen, CTO of Cadence’s IP Group and longtime computing-architecture guru, talks often of “design to the data.” It’s a way of thinking about enabling systems that understands what type of data your system will be addressing and how to optimize your system for power.

These new technological cycles and opportunities coupled with the democratization of technology put enormous pressure on engineering teams but at the same time offer those teams astounding opportunity.

Those who master device integration, low power and mixed-signal engineering techniques, ecosystem and design chain dynamics, and hardware-software codesign will win today and tomorrow.

In the process, they will propel us toward the fifth major era of computing, whatever that may be.

Leave a Reply

(Note: This name will be displayed publicly)