IoT Sessions at 2015 GSA Silicon Summit – Part II

Introduction:

In this second of a two-part article series, we review the IoT afternoon session at the April 15, 2015 GSA Silicon Summit. Part I summarized the morning session and is available here.

MEMS (Micro Electro Mechanical Systems) and Sensors, Shaping the Future of the IoT:

  • Todd Miller, Microsystems Lab Manager, GE Global Research
  • Behrooz Abdi, President and CEO, InvenSense
  • Steve Pancoast, VP, Software and Applications, Atmel
  • David Allan, President and COO, Virtuix Inc.

Todd Miller told the audience what GE cares about for the IoT. One concern is that ~40% of skilled U.S. manufacturing workers will retire in the next five years. That’s a huge challenge for the Industrial Internet, because there’ll be an acute shortage of workers to make the devices/controllers. More outsourcing of high-tech manufacturing to Asia?

Other important challenges include: performance, mitigating cybersecurity threats, scale, and interoperability via open standards. Costs that don’t scale well will limit the value created, Todd said.

GE’s Industrial Performance and Reliability Centers maintain critical asset operations with 6K+ assets in 770 world-wide sites which are monitored 24/7. A wind power site was given as an example.

GE is a founding member of the Industrial Internet Consortium (IIC), an open membership, not-for-profit group of public and private institutions that focuses on:

  • Developing use cases and test beds
  • Sharing best practices, reference architectures, case studies
  • Influencing global standards development to ensure interoperability
  • Building confidence around new and innovative approaches to security
  • Other founding members include: AT&T, Cisco, IBM, and Intel

Miller said that the value of IoT to customers will be huge. Connected machines could eliminate up to $150 billion in waste across industries, he said. Five such industries were cited: Aviation, Power, Healthcare, Rail, Oil and Gas. There are IoT/connected machine benefits provided for each.

GE Global Research provides innovation via “breakthrough device concepts, which become real working devices… from prototypes to low volumes.” A GE MEMS Relay Product Line was established in 2014 with external shipments scheduled for Q4 2015.

Of course, the biggest threat for industrial control is security. In this author’s opinion, the vulnerability of critical infrastructure such as energy and utility is vastly underestimated. A report by Ponemon Institute and Unisys titled “Critical Infrastructure: Security Preparedness and Maturity,” highlights the striking disparity between awareness of cybersecurity risks and the implementation of security protocols in critical infrastructure sectors. More information on this important topic is here.

The Industrial Internet Security Working Committee will establish a security framework to be applied to every technology adopted by the IIC. The framework shall ensure sufficient cyber security and privacy for the various users of the industrial internet. The Security Working Group will also point to best practices and identify gaps. Good luck!


Behroz Abdi characterized IoT as a new form of “Ambient Computing”AlwaysOn with Intuitively Interactive Apps and Services. Another descriptor given was “The Internet of Sensors,” with functions [f(x)] for location determination, activity, time, and environment.

InvenSense was said to be a company that integrates sensors on a SoC, develops algorithms & software, as well as doing systems integration. “The fabless model for MEMS is based on process technology and sensor integration,” he said. The SoC functions from InvenSense often include on-chip building blocks like FIFOs, a digital motion processor, activity classifier, inertial sensors, a tilt sensor, device context gestures, and wake-up sensors.

Abdi said that MEMS technology for a “motion tracking solution” has become more of a software business with 2/3rds of InvenSense’s hires involved in algorithm development for deep learning and software integration.

“Sensors are transformative and are fueling the Internet of Things,” according to Behroz. That’s illustrated in the chart below.

Sensors are transformative according to InvenSense.
Sensors are transformative according to InvenSense.

Wearables are a very promising market for the company. Wearable computing, sports equipment, fitness/activity tracker, virtual reality, head mount displays, extreme sports cam (GoPro camera?), fitness watch and smart pods were cited by Behrooz as wearable IoT products.


Steve Pancoast talked about MEMS and Sensors, Shaping the Future of the IoT. There is a lot of non-digital information processing in the IoT that’s doesn’t follow Moore’s law, he said. That includes: sensors, RF and passive/discrete components.

In particular, edge sensing nodes are and will be a large part of the IoT, Steve said. Some examples are provided in the graphic below:

Edge sensing nodes are a major part of IoT.

Those IoT edge nodes have a broad range of applications and that diversity mandates the following:

  • A very broad portfolio of low power MCUs and MPUs
  • A diverse portfolio of easy to use, secure, relevant wireless products
  • A complete solution where the system software becomes a key differentiator

A complete IoT solution from Atmel will usually include: Wireless Connectivity, Security/Privacy functionality, Low Power Embedded MCUs and MPUs, Sensors, and Software/Tools

IoT Communications Topology will be very dependent on the industry vertical, whether the end device is in a building/home or in the field (different network connectivity) and what type of gateway (if any) is needed to connect things/ endpoint devices to the Internet. This is depicted in the chart below:

The IoT communications topology is depicted in this diagram.

Atmel SmartConnect was said to bridge the gap between embedded hardware/firmware/ software developers and backend services/software developers, as shown in the illustration below:

What does it take to make IoT a reality?

Sensor’s big technology ally was said to be “contextual computing,” which will determine “Where, When, Who, How, and What.” Atmel says that “Contextual computing will be the driving force behind the next wave of new technologies.” We’ll see…

A very comprehensive IoT layered security diagram is shown below. It illustrates each protocol stack layer and the corresponding security function/protocol. The key point is to provide critical security for each and every IoT edge node.

SmartConnect IoT layered security solutions is depicted

IoT endpoints were said to be “a natural fit for Atmel MCUs.” Steve stated that Atmel has:

  • Complete Range of Processing Cores: ARM Cortex M0+, M4 & A5/A7 MPUs
  • Industry leading Low-power, SmartConnect Solutions: Wifi, BT/BLE, 15.4 coupled with Cloud Solutions
  • Sensor Hub Solutions & SW with wide industry support
  • Large selection of Robust IoT Crypto Solutions & Security software

In closing, Steve told the audience to “dream big about IoT” as he showed a photo of a fish wearing what looked like an IoT harness with embedded sensors.


David Allan was very poised as he delivered his closing conference presentation by welcoming attendees to the “Second Machine Age.” Hello: smart house, connected car, connected person, and even a connected cow!

After quoting Broadcom’s founder & CTO Henry Sameli, PhD that “Moore’s law is coming to an end” David boldly claimed that “Moore’s Law doesn’t matter!” He believes that the rise of distributed computing makes transistor densities and processor clock speeds less relevant than before.

[Coincidently, an article in the Economist magazine made the same point: “With the rise of cloud computing, the emphasis on the speed of the processor in desktop and laptop computers is no longer so relevant. The main unit of analysis is no longer the processor, but the rack of servers or even the data centre. The question is not how many transistors can be squeezed onto a chip, but how many can be fitted economically into a warehouse. Moore’s law will come to an end; but it may first make itself irrelevant.]

David cited Google’s work on “MapReduce: Simplified Data Processing on Large Clusters” as being relevant for the IoT.

The key characteristics of MapReduce are to:

  • Scale down: to minimize number of nodes
  • Scale up: to maximize number of nodes
  • Assume failures are common
  • Move processing to the data (data locality)
  • Avoid random access

Mr. Allan defines IoT as wireless sensor networks connected to the cloud, thereby harnessing the power of distributed computing.

[But not all the wireless sensors/IoT endpoints will directly connect to the cloud. Many will communicate with a local controller/ gateway or to each other. For example, the Intelligent Proximal Connectivity (AllJoyn) is a collaborative open source project of the AllSeen Alliance that aims to enable apps to connect, control and share resources with other nearby apps and connected smart things.]

Deep Reactive-Ion Etching was said to be important for MEMS, but also for advanced 3D wafer level packaging technology, which might be used for IoT sensors and endpoints.

MEMs are used a great deal in mobile devices, as illustrated by the schematic diagram below:

MEMS in mobile devices from Virtuix.

Mr. Allan is quite concerned about IoT standards and non-conformance of sensors to performance specifications. David wrote in an email:

“Yes, we see a great need for standards, in particular harmonized performance standards for sensors. Sensor devices which—according to datasheets—have equivalent performance, often differ in reality.

For example, after the iPhone 5S switched from a three-axis STMicro LIS331DLH accelerometer to a seemingly equivalent Bosch BMA220 part, many applications (mostly video games) suffered a loss of accuracy of as much as five degrees! Some magnetometers didn’t perform according to specs.

In the future, we’ll decide which part to populate after extensively testing our production boards. Clear performance standards would make this decision possible up front.”

Somewhat whimsically, David asked the audience: “What will the second machine age look like?” His futuristic answer:

“Our new machines will augment human desires…”

  • Immortality
  • Omniscience
  • Telepathy
  • Teleportation

Personally, I’ve been waiting for teleportation since I watched the original Star Trek in college. “Beam me up Scotty.” Over and out….

Till next time……………..

0 thoughts on “IoT Sessions at 2015 GSA Silicon Summit – Part II

  1. Once again, excellent summary, kudos to Alan. DRIE is indeed one process that made its way into 3DIC for TSV’s riding on the coattails of its legacy and valuable experiences from MEMS fabrication. I would argue that the industry needs lower cost and faster processes to fabricate 3D integration with MEMS and SoCs for IoTs for the billions of volume forecast.

    Whereas the comment “defines IoT as wireless sensor networks connected to the cloud, thereby harnessing the power of distributed computing” is largely true, there are many use cases where edge-computing is absolutely necessary for critical smart city infrastructure, for example.

  2. Very comprehensive follow-up to part 1, Alan. Thanks for sharing your insight and research regarding this conference with the Viodi View readers.

    The comments from Mr. Allan regarding the omniscience reminds me of a comment made at the TIA conference on connected/autonomous vehicles. One of the panelists commented that it is probably appropriate to start having ethicists at these sorts of gatherings (in that context the question was about how to program ethics into the self-driving cars).

    1. Thanks Ken. I actually had planned a lot more editorial comments, especially about the misguided notion there would be ONLY a few standard IoT SoC’s. However, the article was already too long. Those who might be interested in pursuing a consulting arrangement with me are invited to inquire by emailing me: alan@viodi.com

  3. From April 27 WSJ on “Connecting Everything”
    “…in the not-too-distant future, the question won’t be what devices are connected—it will be what devices aren’t connected.

    Gartner Inc. estimates the number of connected devices in use will reach 4.9 billion this year. By 2020, there will be 25 billion. With smaller processors, wider networks and more sources of energy, it won’t be long before everything around us becomes a computer, from our headphones to our chairs.

    Of course, this comes with an element of risk. Connected devices open us up to possible security issues and intrusions into our privacy. But if we’re able to strike a balance between caution and convenience, the spread of connected devices will have a profound impact on the way we do just about everything.”
    http://www.wsj.com/articles/nest-ceo-tony-fadell-on-the-future-of-the-internet-1430104501

  4. Thank you, Alan, for the excellent reporting. These two articles were among the most informative that I have read about IoT in the last year.

Leave a Reply

Your email address will not be published. Required fields are marked *

I accept that my given data and my IP address is sent to a server in the USA only for the purpose of spam prevention through the Akismet program.More information on Akismet and GDPR.

This site uses Akismet to reduce spam. Learn how your comment data is processed.