By Françoise von Trapp, Editor-at-Large, IPSO Alliance
This was my first IoT Slam, as well as my first-ever virtual event. I spent eight-straight hours attending as many real-time events as I could fit into my schedule (some overlapped), and came away with an abundance of information that I will attempt to boil down into what I consider to be the key takeaways.
Full disclosure: My primary area of expertise is in the semiconductor and MEMS sensor manufacturing sector of the Internet of Things (IoT), so a conference that focused mostly on data was a new experience for me. I attended to learn more about how the IoT will be monetized, and, more importantly, what’s being done to address security and privacy issues.
While semiconductors provide the underpinnings of the IoT, it’s clearly NOT where the revenue will be realized, other than in increased volume. This is because, as James Turino of Redwood Capital Group pointed out, one of the IoT growth drivers is the decrease in infrastructure costs of MEMS, sensors, storage, servers, etc., and the ability to use off-the-shelf standardized components. He also said the cost of sensors needs to drop more. He noted that of four IoT business models—classified as hardware, connectivity, software, and managed services—software stands to have the highest gross-margin profile, while hardware will have the lowest. However, all of them are critical to IoT success.
The overarching message of IoT Slam is that the IoT is changing the way the world does business. Sam George of Microsoft Azure said that the IoT is a “new normal,” just like the Internet, the cloud, and mobile products have become, and the early movers will benefit most. Turino also noted that the “return on investment of the IoT is large, measurable, and immediate.”
Monetizing the IoT is about analyzing the data. But agreeing how best to do that is still being debated.
As Bryson Koehler of the Weather Company pointed out, every day there’s a new IoT sensor coming online providing new connected data gathered from a swath of utilities that allow them to reach a broader audience. He views IoT data as an asset of crowd-sourced information. “If we use it for good, we can help people, businesses, and governments make smarter decisions,” he said. For example, real-time weather data could help prepare for major weather catastrophes, and prevent incidents like the Atlanta ice storm from happening.
Another example of the importance of real-time weather data for aviation applications. Thanks to IoT-based data coming off aircraft sensors in real time, it’s possible to reroute planes to avoid turbulence. This saves cost in terms of maintenance inspections and flight delays due to airport congestion. Koehler says data show savings of over $4M in fuel thanks to decreased taxiing time.
Koehler also talked about the value of this data: It helps businesses understand how weather impacts people’s behavior. For example, they’ve noted a 24% increase in sales after activation of weather-based triggers. By tying data to historic, current and foreign weather conditions, companies can understand sales patterns during weather conditions, then use the information to manage inventory.
Steve Blackwell of Emerson Network Power noted that one challenge in taking all the disparate data and turning it into an aggregated, usable set of information has to do with the fact that we are dealing with the Internet of legacy devices. How do we make devices never intended for the IoT into IoT devices?
In the big data vs. fast data debate, Emil Berthelsen of Machina Research says it’s really about a hybrid of the two. Fast data is allowing new applications to exist. Without the ability to analyze, capture, and store data at a fast pace, we wouldn’t be able to support applications we use now. For example, in a connected car, latency is critical. But the ability to analyze real-time data alongside historical data provides an aggregated view, he explained.
Patrick McGarry of Ryft concurred with Berthelsen, noting that the value of IoT data is in when it was captured—not two days, two weeks, or two years from now. Additionally, processed data is more valuable than raw data. IBM’s Sandy Carter compared its value with oil: The more refined it is, the more valuable it is.
This push for hybridized, fast, processed data is building a case for a future that relies less on the cloud, and more on computing on the edge, noted Scott Amyx of Amyx McKinsey. He talked about processing data in real time on our devices because there is no time to make a round trip to the cloud. This is enabled by billions of sensor devices and high-performance chips that can be autonomous. He also talked about mesh networking that uses a point-to-point connection to transfer sensitive data. To illustrate the need for real-time data analytics, he used the example of a driverless car that gathers information about the surroundings of the car, for collision avoidance. Processing must take place at the car level.
Ryft’s McGarry explained that to analyze data at the edge where it is generated requires a combination of central processing units (CPUs) and field programmable gate arrays (FPGAs). This is because the CPU environment alone is reaching the end of Moore’s law, and is not capable of the parallel processing required to produce real-time data. FPGAs are the only architecture capable of highly parallel processing. Intel understands this, noted McGarry, and that’s why they bought Altera, an FPGA company.
Security/privacy was the second area I focused on in choosing which presentations and panels to attend during the IoT Slam.
As noted by IPSO Alliance’s Geoff Mulligan, there is a difference between security and privacy. Security is a technology issue, such as data encryption. Privacy is a policy issue, such as who owns your data, or who should be able to access it? One potentially scary issue Mulligan pointed out was a de-aggregation problem, which he described as being able to determine an individual’s location simply by accessing three random data points throughout the day.
Christian Légaré, of IPSO Alliance member company Micrium, said the IoT is more than a “thing” or “the cloud.” It transforms a product business into a service business. He also talked about the “crash of domains”: Trying to put Internet technology onto an end device is like “putting a square peg in a round hole.” Lots of IoT security is focused on the cloud and data, not the end device. This is the problem that the IPSO Alliance is addressing, with its focus on promoting open standards for Internet protocol for smart objects.
The key to identity management, said Mulligan, is for each device to have its own identity, so we know where the piece of data came from. To do this requires a harmonization of protocols, which is why IPSO Alliance promotes open standards and open protocols. “IP stands for internet protocol. We think IP should be on everything,” he said. “By creating a standard, it allows us to work on harder problems like privacy, security, and life-cycle management.”
Brian Witten of Symantec noted that IoT device security requires transparency, government regulation, and infrastructure updates. By transparency, he meant that people have a right to know how much security is built into a device without reverse-engineering it. Part of the reason that security isn’t built in is because people don’t know how to do it. Just because devices need to be low power, small and low cost are not valid excuses for not building in security.
Atif Hussein of Intel offered three steps to success for IoT security: harden the devices; secure the communications; manage and monitor, and monitor again.
IBM’s Chris O’Connor noted that ecosystem partnerships are key to creating an IoT platform that is secure from chip to cloud.
I’ve barely scratched the surface of the topics that were discussed throughout the day at IoT SLAM 2016. In sharing these lessons I’ve learned, I hope I’ve inspired you to register for the event, which has all the presentations available for on-demand viewing for the next three months. Check it out at http://iotslam.com/.