IoT news of the week for April 12, 2019

More layoffs at GE Digital: The software and digital transformation arm of GE is laying off a chunk of workers in California, further reducing the staff of the beleaguered business unit. The company claims this layoff isn’t related to the company’s eventual spinoff, but is because of “commercial demands.” I wonder if it’s a lack of commercial demand for GE’s IoT platform. The industrial giant is credited to bringing mainstream attention to the industrial IoT, but it’s marketing efforts were expensive and may have outpaced overall demand.  (CRN)

Google’s Anthos is good for the IoT edge: This week, Google held its annual cloud computing event, and for infrastructure nerds like myself it’s always fun to see what the Goog comes up with. This year’s two big announcements from an IoT perspective were Google Anthos, an open-source cloud management tool that lets someone run computing jobs across multiple cloud providers and on premise. For those that don’t speak cloud, that means you can build something to run across several computing platforms without worrying about vendor lock-in. The attention to on-premise computing is good for companies that want to host some parts of an application at the edge. All of this is made possible by the use of containers and Kubernetes to manage the containers. Still up in the air for me is how Google plans to handle security in a multicloud world and how it allocates resources between clouds. (Could storage be on AWS while compute is handled in Azure? This seems impossible.) Also, how will this affect pricing? And will it allow companies to arbitrage price differences in different clouds? Because if so, this is a big deal. Anyhow, there’s a lot to wonder about Anthos. (Geekwire)

Google’s Cloud Run gives Amazon’s Lambda a run for its money: Also at its Cloud Next event, Google announced Cloud Run, a serverless infrastructure that lets developers spin up containers as needed without having to worry about the underlying physical infrastructure. For IoT, where a developer might only need to compute something once an hour, serverless architecture makes a ton of sense. Why pay to run a server for 59 minutes of what is essentially downtime? So far, the reception to Cloud Run has been positive, with devs stating that it’s easier to work than Amazon’s Lambda serverless platform, but we’ll see how long that’s the case. (Geekwire)

One year into Microsoft’s $5B IoT pledge: A year ago, Microsoft said it planned to spend $5 billion on IoT and edge computing, but didn’t go into much detail as to what it would spend that on. It has now published a blog post showcasing some of the customers it has helped connect. The likes of Starbucks (connecting its coffee-making equipment), Walmart (connecting its HVAC and refrigeration gear), and Chevron (connecting heat exchangers for predictive maintenance) all are sending data to Microsoft’s Azure cloud. Surprisingly, Microsoft also touted its work with Volkswagen and BMW, both of which have cloud deals with Amazon Web Services, a Microsoft Azure rival. (Microsoft Blog)

Security is about compromise: It’s impossible to make something totally secure and still get use out of it. That’s why security researchers recommend that businesses continuously think about security and assess risk models. But as connectivity expands throughout more industries, CIOs and CISOs aren’t actually keeping up. What surprised me was just how far behind they really are. A study on resiliency by security firm Tanium says that four out of five CIOs and CISOs have discovered that a critical update or patch they thought had been deployed had not actually updated all devices, leaving the business exposed as a result. There are two problems here. One is the lack of visibility that CIOs and CISOs have into devices on their network, and the other is that necessary patches aren’t getting done. Sometimes I think we need to put a moratorium on connecting equipment until everyone gets an education in basic IT security. (Tanium)

Intel needs the edge, but does the edge need Intel? VentureBeat has an interview with Jonathan Ballon, the head of Intel’s IoT efforts, in which Ballon talks about how more than half of data is being created in the physical world and for jobs that require extremely low latency. Companies, therefore, don’t want to move that data to the cloud and then back. That means demand for edge processing will rise (see the Flex Logix story above). The interview does a good job proving that Intel understands the trends, but I wonder whether Intel has the silicon required to succeed — or dominate — in this space. It purchased Movidius as well as Vertex.ai, both of which boost Intel’s credibility in edge processing. But Intel built a culture and manufacturing business around a monolithic x86 processor, and I don’t know how that fares in a time where beefy, power-hungry chips are needed for fewer jobs. (VentureBeat)

Armis gets $65 million for IoT security: Armis, which provides security software that tracks connected devices on a corporate network, has raised $65 million in funding to help it broaden its customer base. I’ve written about Armis before, and the approach it has makes a lot of sense for connected devices, which can be too resource-constrained or too numerous to run traditional security software on the device being monitored. It’s part of a wave of fundings in the IoT security market, and with this round, Armis brings on Carl Eschenbach, VMware’s former president and COO, who is now a partner at Sequoia, to its board. Armis expects Eschenbach to help the company increase sales among enterprise buyers. (CRN)

Lessons on using Kubernetes for IoT: This is an older blog post, but I think it’s newly relevant given Google’s cloud announcements this week around Cloud Run (a serverless computing framework) and Anthos, a multicloud computing platform that relies on Kubernetes. Bosch developers lay out what they learned using Kubernetes, including how to use a number of helpful tools, but also caution developers that it’s not something they can “set and forget.” I actually hope they update this resource in a few months because as they say in the blog post, this is a rapidly evolving community, and I would love to know what new tools are available. (Bosch)

Don’t panic, but yes, people can listen to your Alexa utterances: At first, this story scared me, but after reading it I realized that there’s actually nothing to truly freak out about here. The story says that human employees at Amazon are listening to your Alexa, but if you read through it turns out that human Amazon employees listen to a small fraction of the words you day after someone says “Alexa” or another of the Echo’s wake words. You can also opt out of sharing your utterances with actual people if you’d like by adjusting the privacy settings in the Alexa app. Most of us were aware that anything heard after Alexa wakes up gets sent to Amazon, but now we know that sometimes a real person might hear that request. Amazon says the utterances are not associated with users’ names. Apple also uses real people to improve its voice services, and Google does as well. Although in the case of Google, voices are distorted to further protect user privacy. (Bloomberg)