How to think about AWS Greengrass and the internet of things

The cloud can’t always compete with local processing. (Slide courtesy of Amazon.)

Connectivity isn’t a constant and it’s not cheap. While consumers might be willing to trade the reliability of analog systems such as a land line for the slightly less reliable cell phone, enterprises and industrial conglomerates are not.

That’s why Amazon Web Service unveiled a new product in Las Vegas last week called Greengrass aimed at linking edge devices to Amazon’s cloud-based web services. A device running AWS Greengrass core will likely be a heftier gateway device that sits inside a building, a truck or a home and connects with other sensors or devices in the building.

Greengrass allows companies to track state and offer on-device analytics. Plus, it lets devices that use the AWS IoT Device SDK communicate with the device that hosts the AWS Greengrass core without connecting back to the cloud.

Amazon’s Werner Vogels, who is the CTO of AWS, explained it well in his blog on the launch:

“As it turns out, there are three broad reasons that local data processing is important, in addition to cloud-based processing. At AWS we refer to these broad reasons as “laws” because we expect them to hold even as technology improves:

  1. Law of Physics. Customers want to build applications that make the most interactive and critical decisions locally, such as safety-critical control. This is determined by basic laws of physics: it takes time to send data to the cloud, and networks don’t have 100% availability. Customers in physically remote environments, such as mining and agriculture, are more affected by these issues.
  2. Law of Economics. In many industries, data production has grown more quickly than bandwidth, and much of this data is low value. Local aggregation and filtering of data allows customers to send only high-value data to the cloud for storage and analysis.
  3. Law of the Land. In some industries, customers have regulatory or compliance requirements to isolate or duplicate data in particular locations. Some governments impose data sovereignty restrictions on where data may be stored and processed.

Today, we are announcing the general availability of AWS Greengrass, a new service that helps unlock the value of data from devices that are subject to the three laws described above.”

This is a good move for Amazon that extends its cloud computing efforts beyond its own data centers and validates the concept of edge computing.

Cisco has been talking about this for at least five years as “fog computing” while Microsoft recently launched a similar edge-to-cloud program called Azure IoT Edge. Other big IT shops are offering similar products, such as HPE’s ruggedized Edgeline servers that have up to 64 Xeon cores for data analytics or Dell’s efforts to build a community around EdgeX Foundry.

There are a few things to keep an eye on from a business perspective.

First, is this idea of Amazon moving beyond the cloud. Fintan Ryan, an analyst at RedMonk, points out that for the last three years Amazon has been embracing on-premise computing for customers. For example, it offers a storage service called Snowball that let companies back up huge amounts of data on physical hard drives, which Amazon then takes to its own data centers using a truck. Once the data is sent to Amazon’s long term storage the drives are delivered back to the customer for a refill.

Ryan views this sort of hybrid cloud/on-premise offering as essential for some of the tasks companies need for industrial IoT. If Snowball is a way to reliably and cheaply move vast quantitiesof data around (and Glacier is a cheap place to store it over the long term), then Greengrass is a cheaper way to take action on relatively short-lived data that’s only valuable for a brief moment in time. When that moment is happening, it’s running on Greengrass. When the moment is past, it gets shipped to Glacier using Snowball.

Later, if a company wants set up a model of the factory at a given point in time based on that data, it has the tools.

The second thing to watch here is the pricing. Amazon prices many of its services based on consumption. But when you are installing software on edge devices, the model is generally based on a license per unit. Look to see how Amazon adapts pricing for that market, or if the market adapts to more of a consumption model. For now AWS is charging a per device fee on the first 10,000 devices, but large-scale customers will likely have many more devices than that.

Finally, look for startups offering tools to make the management between Amazon’s serverless computing platform called Lambda and Greengrass work. There’s a huge opportunity for better monitoring of these systems and making it easier to set them up.

For example, Ayla Networks, a startup that offers a hosted IoT platform to customers such as Fujitsu and Hunter Fan Co., built its cloud services on AWS. Now, it can extend its offering to industrial minded clients who need a gateway to manage local processing and events.

Bill Podrasky, VP of global business development at Ayla, explains that while some customers can build a platform themselves on AWS’ platform, there are plenty of customers that want something much easier to consume. He too, is curious about that pricing.

Stacey on IoT | Internet of Things news and analysis

Internet of Things News of the Week, June 9 2017

Police in Wales have arrested a man based on automatic facial recognition.

Here’s a summary of interesting analysis and news about the Internet of Things from the past week. Get this summary in your inbox every Friday when you subscribe.

Worrisome news on the NB-IoT and LTE-M front: The cellular carriers have been touting two standards for low power wide area networks. LTE-M has higher data rates, consumes more power and is in the middle of rollouts in Europe and the U.S. BB-IoT has a smaller payload and is reportedly cheaper. The idea is that instead of using a technology such as Sigfox or LoRA, a device manufacturer needing a cheaper low power network option would stick with cellular. The reality is proving to be different. Both articles have European operators discussing some of the challenges associated with the cellular LPWAN tech so far. Price, security and interoperability concerns are apparently slowing adoption. (Light Reading, The Register)

Can animal cognition teach us about building distributed IoT networks? Uh, maybe? My mind was blown as I read this article about spiders and octopuses and how they think, having evolved distributed and external thought processing techniques. Then I started thinking about how the new EdgeX Foundry distributed middleware project for industrial IoT has an octopus as its emblem, and wondered if we could learn more about building networks from how these animals think. It’s not that far out there. (Quanta Magazine)

Annie get your contour makeup: Looks like police in Wales have arrested a man based on automatic facial recognition. The case provides a scary glimpse into the future where the government has a record of our faces and can apply it to people surveilled in public places. When it comes to catching criminals people may applaud this use case, but it could become a  tool for censorship or even extortion. The way out may be ever odder makeup and hairstyles to stymie the algorithms. (Ars Technica)

Speaking of surveillance: If you have a smart speaker in your home, such as an Amazon Echo or a Google Home, you may want to read this overview of how each of these products protects your privacy. Or doesn’t. (CNET)

How to hack a pacemaker: Dick Cheney was ahead of his time in deciding to break the wireless functionality in his pacemaker. Security researchers who looked at seven pacemakers made by four different vendors found 19 vulnerabilities shared by most of them. (Quartz)

This one is for the makers and prototypers: Microchip has launched a microcontroller with an integrated graphics processor.  So if you have a device that powers a larger screen, but is still relatively dumb, this might be the chip for you. (

Want to pitch the FTC on your privacy technology? The call for ideas and papers to present at the FTC’s PrivacyCon event next year is now open. Topics include: How do companies assess consumers’ privacy preferences; What are the greatest threats to consumer privacy today; What are the costs of mitigating these threats; How are the threats evolving; and more. (FTC)

Privacy may become a competition issue: The large data repositories companies hold on consumers and their ability to use that data to erect barriers to entry for newcomers may result in new thinking around antitrust. This is a concept we’ve touched on before in this newsletter, and apparently, the Germans are thinking about it too. (Wired)

IBM made a 5-nanometer chip: For all the talk of Moore’s Law fading away, the chip industry is doing back bends while roller skating to keep it going. The effort of cramming more transistors onto a chip is getting technically more challenging to build thanks to the laws of physics, but that hasn’t stopped IBM from unveiling a 5 nanometer chip this week. Big Blue and its partners in this endeavor say the soonest such a chip might make it to market is 2019, but go ahead and learn about it now. It’s a pretty big deal, even as the idea of massively powered processors seems somewhat quaint in our mobile, modular world. (Wired)

Your offline shopping is heading online: Two staples of online shopping, the ability to track what you are buying and dynamic pricing, are making their way into the physical world thanks to the awesome powers of data collection and analysis. This MIT Technology Review story explains how Google can use your credit card and Maps data to track if you have made purchases in a physical store after seeing an ad, while the Guardian worries about dynamic pricing based on your shopping history and demographic data. (MIT Technology Review, The Guardian)

Want to know what’s new in HomeKit? Watch here for new features, including data on the addition of speakers as a device type. Wheeeee! (Apple)

Stacey on IoT | Internet of Things news and analysis

All about Apple’s HomePod

The Apple HomePod. Image courtesy of Apple.

Apple’s disclosure of the HomePod, a connected speaker and personal assistant, drove much of the IoT news this week. However, research from Pew on how rapidly people are becoming connected and the lack of transparency about how our data is used might end up being the story with real legs. Kevin Tofel and I discuss both this week, along with some Wink news, how he feels about the Google Home and a brand new purchase I made.

To continue with the HomePod theme, I spoke with three different people to get a sense of how voice affects adoption of smart home technology, what the HomePod could mean for HomeKit adoption and what another voice-activated speaker means for privacy. Scott Harkins of Honeywell, Adam Justice of ConnectSense and Nuala O’Conner of the Center for Democracy and Technology joined me for the discussion.

Stacey on IoT | Internet of Things news and analysis

Visybl wants to bring easy IoT to small biz

Visybl’s goButton’s use Amazon’s Dash buttons to make boring office tasks easier.

Here’s an IoT crowdfunding campaign to watch!

Normally I’m not that into crowdsourcing campaigns because, like those pictures of fast food burgers, they tend to disappoint in real life. But Visybl has a cool idea to take Amazon Dash buttons and repurpose them for small businesses…taking a concept I’m meh on, to something that seems like a nice way to make life easy for office managers.

Visybl is doing a Kickstarter campaign that launches later today that will offer customizable Dash buttons for small business owners or even individuals. For example, if you send a lot of FedEx packages, just program the button to notify FedEx that a delivery is ready for pickup. With one press you set the process in motion, avoiding a phone call.

You could do something similar for scheduling a service call for the copier or use it to buy refills for the coffee maker.

The buttons, which should ship in October, are $ 44 for one or $ 400 for a package of 10. They tie into Visybl’s cloud service and use a graphical user interface to help non-tech users set up the buttons. Amazon Dash buttons are only $ 20 but you have to program them yourself. There are plenty of people who will not want to do the backend developer work to make that happen.

These could even be used to set things in motion in your smart home, although I think there are plenty of other excellent button options out there.Visybl has been in business since 2015 and its primary product revolves around beacons and location services. Offices buy a gateway device and use the Visybl beacons to track things around the office. The company describes itself as Tile for the enterprise.

Based in Germantown, Maryland, Visybl has funding from the Maryland startup Fund. It makes money on larger purchases of its devices. For the goButton, it expects that customers wanting to buy more than 100 will also need provisioning, configuration and management that Visybl can provide. But for the smaller customers, the cost is just the one-time purchase of the device.

In my old office, I would have killed for a button that I could press to shame my colleagues into cleaning up the communal kitchen after themselves. That’s when the IoT gets really good.

Stacey on IoT | Internet of Things news and analysis

Hackers aren’t just a security issue, they are a safety issue

Professor Ross Anderson explaining how IoT security becomes a safety issue.

Think about your first car. If you’re like many of us, it was probably used, either purchased or handed down from parents or maybe even siblings. My first car was 14 years old. Now, think about how poorly secured most 14-year-old electronic devices are. The success of WannaCry capitalized on precisely that lack of security for hospital machines and operating systems.

When it comes to embedding tech into everyday machines and critical infrastructure, security is no longer just about privacy of data, it’s a safety issue. To build a connected medical device without contemplating how to prevent someone from hacking it is like building a bridge without a civil engineer and load testing. The lack of either could result in the loss of life.

That’s been missing from a lot of the conversations about IoT security, but an excellent paper from the University of Cambridge looks at security for connected cars, medical devices and the energy grid as a safety issue.

The report covers the role of regulators in our technocentric world, how liability might be shared among manufacturers, consumers, insurers and regulators, and ways to implement security over a 30-year life span. Alone, each of these topics is worth several papers, which is why I found the report such helpful reading. It covers all of these to enough depth to help interested parties dig into the issue, rather than say, “It’s hard,” and move on.

The paper offers up liability as the impetus for secure infrastructure and discusses where European laws (this report was commissioned by the European Commission) fall short. For example, product liability can invalidate the ubiquitous end user licensing agreements that people click on with every software download, but it cannot protect against unforeseen harms.

And when dealing with the interest, electronics and a network, we will encounter the unforeseen. Additionally, laws that protect against insecure and unsafe products do not cover services. For many companies, connected devices are about selling a service, not selling a physical device.

There are an additional half dozen interesting points for discussion in the paper, plus a historical perspective on how regulation developed around the railroads and cars. It provides grim parallels on how things might unfold for connected device security in the 21st century.

If all of this sounds crazy and over-the-top, go reread the essay I wrote about hospital security. The CISOs I spoke with for that story were most worried about their connected devices doing physical harm to a patient, not about records getting stolen. We’re already living in a world made less safe by connected products. Now we have to admit this and solve the issues technology has wrought.

Just like we have done before.

Stacey on IoT | Internet of Things news and analysis

Internet of Things News of the Week, June 2, 2017

Dewalt’s connected construction site plans include Wi-Fi and asset tracking. Image courtesy of Dewalt.

Here’s a summary of interesting Internet of Things news from the past week. Get this summary in your inbox every Friday when you subscribe to the newsletter.

Intel’s Compute Card is worth a look: There are many technical limitations that long-lived devices present. For example, a connected oven might live in your home for 15 years while the electronics inside die after five. So how do connected device makers continue upgrading or replacing old tech in cars, appliances and other products? Modular housing for electronics is the solution most companies are discussing whether it is Jenn-Air or Audi. And Intel just showed off something that may fit the bill. (Although it seems too powerful for many embedded use cases.) The chipmaker unveiled a Computer Card at Computex that slots into machines and contains the computing and memory a device needs. It’s a long way from this demo to production, and most of the demos seemed to revolve around turning a monitor into a full-fledged computer, but a mass market upgradeable compute card has a lot of potential value.  (Engadget)

Europe is looking at expanding privacy regs for IoT: The EU already has some privacy regulations around cookies and data collection on web sites, but it now aims to expand them to non-web devices. You can read the amended proposal here.

A robot learned how to pick up real things in virtual reality: This story is kind of mind-blowing for me. It’s about a robot that was learning how to pick up 3-D objects (which is easy for humans but tough for robots) and it did so by picking up virtual versions of the object. The robot was trained in virtual reality with coded models and, by picking up millions of things, the robot taught itself how to pick up real-world objects. This has huge implications for training robots on delicate tasks that might involve people, such as helping hold onto elderly people as they walk. This isn’t an area where a robot can afford to make mistakes with real people, but virtual models of humans may be all that’s needed. Honestly, I’m not 100% sure what this has to do with IoT but it’s a pretty big development. (MIT Technology Review)

Everybody is getting high on home Wi-Fi: After what feels like years of complaining about lame home Wi-Fi, it seems the big companies are responding. Comcast invested in Plume and now Qualcomm is building a mesh Wi-Fi system while Samsung this week unveiled its new mesh-Wi-Fi routers. Mesh isn’t a panacea but it does solve some of the problems user have as they try to extend their Wi-Fi networks throughout a home. The latest Wi-Fi tech promises better coverage and more user-friendly software for must-have features like guest networks and device data. (CNET, Engadget))

Another big risk for connected devices: We spend a ton of time discussing device security, but less time discussing what happens as non-tech companies try to open up their services to consumers with mobile apps or data-sharing. Without the proper security practices these apps are dangerous gateways into customer data and can even provide access to back end systems. Even if you encrypt the data on the way to a database, you have to make sure it is secured.  (eWeek)

Softbank invests in OSIsoft: Softbank, which bought ARM last year in a bid on the internet of things, has made an investment in OSISoft, buying stakes from Kleiner Perkins, TCV and Tola Capital. OSIsoft makes software that helps manage data coming in from industrial processes. More than 1,000 utilities, 95 percent of the largest oil and gas companies and other large industrial conglomerates use its PI system software. The company touts itself as a big industrial IoT player, but I am curious how its proprietary system weathers the up-and-coming generation of efforts to make the industrial IoT more standards-based.

Dewalt is building an IoT platform: Tool maker Dewalt is going all in with an IoT platform for construction sites. Dewalt plans a mesh Wi-Fi network product for construction sites and will include services like asset tracking and data analytics coming from the tools and workers. (Dewalt)

Google Home and Nest are two different companies (and they will remain so): So sayeth Ruth Porat, CFO at Alphabet. Speaking at the Code conference this week, she said, “Our view is that Home is pursuing one path and Nest is looking at a different set of products and go-to-market strategy.” All this does in my mind is cement this idea that Nest may soon find itself sold to the highest bidder. Meanwhile, Google will continue to move forward with the operating systems and infrastructure associated with connected devices, as opposed to the hardware itself. (ReCode)

The internet of things will eat your lunch: Everyone in business spouts off about digital transformations, including me. The idea of a lot of cheap data that can be quickly and inexpensively analyzed to provide insights is a big game changer for every business. But while executives recognize this, they apparently have less faith in their digital intelligence than ever. PwC quizzed executives on this topic in 2007 and 2017 and the results were pretty grim. (Harvard Business Review)

So many new cameras! This week Nest launched its 4K connected camera that offers facial recognition and up to 12x zooming abilities. The camera will cost $ 299 and will have some nice features such as the ability to notify you when it hears a dog bark.  Meanwhile, Honeywell also launched a connected camera for the home as part of its Lyric line of devices. The Honeywell camera is $ 119, and seems fairly basic with two-way audio and night vision. (Nest, Honeywell)

Would you buy P2P electricity? I have so many questions about Drift, the startup profiled in this article. It is basically providing customers energy by trying to use excess capacity from other customers who it pays to reduce their demand. It can also buy electricity on the spot market when needed. The idea is that it can offer lower prices thanks to micro-generation and awesome algorithms. I am intrigued, but it also reminds me a bit of Enron. Plus statements from the CEO like, “We flip the priority list over from what a utility does…[which] is to keep lights on. It seems after 135 years, we should be beyond that,” have me a bit worried. (Quartz)

Stacey on IoT | Internet of Things news and analysis

The Essential Phone and IBM Watson

The Essential Phone

On this week’s Internet of Things Podcast, Kevin and Stacey discuss the Essential Phone from Andy Rubin’s new company that features teases a new connected home assistant. Kevin and I discuss what we know, and Kevin also explains a surprise decision he’s made regarding Google (here’s the link he mentions during the show). We also discuss Apple’s potential AI chips, ARM’s new designs and a way to add capacitive touch to wood! Finally, I share my WeMo dimmer switch thoughts.

Bret Greenstein, VP of Watson IoT for Consumer Business at IBM, is our guest. He shares a bit about Watson and IoT, but his biggest service might be his help breaking down how analytics, machine learning and AI all relate.

Stacey on IoT | Internet of Things news and analysis

Haystack Technologies: Building a network for the edge

Details on the way DASH7 is structured.

Wander around any tech event and the topic of edge computing is bound to come up. After a decade or so of preaching the wisdom of clouds, the internet of things is forcing companies to embrace the edge.

Dive a little deeper and you realize that many of the edge promoters are selling boxes that sit between a network of sensors and the cloud doing some form of intermediary processing and then sending the data along. But what if the intelligence in the edge really was at the edge? In the sensors themselves?

That’s the pitch that Pat Burns, the CEO of Haystack Technologies, has. He believes that the current model of edge computing can’t scale with the growth of sensors, and the tech world would be better served by data that stays on a sensor until it’s needed.

The idea is that instead of a device always transmitting state to a gateway or the cloud, instead it would wait until it’s queried. Or, if a threshold was reached it would send the necessary data along. But sending the data doesn’t have to rely on direct transmission. The software Haystack uses is set up as a mesh and can run on devices with as little as 30 kilobytes of memory.

His company makes a software layer that sits on top of any radio that can transmit in the sub-gigahertz range. The software layer is part of a standard called DASH7. I’m not a huge fan of proprietary software stacks, but the industry is currently searching for ways to build easy low power wide area networks for the IoT, and Haystack’s tech can offer that.

Because DASH7 works at sub0gigahertz range, the data can travel farther. Burns is currently bullish on running Dash7 over LoRA radios, but it can also work over many others. To me the most interesting element of my chat with Burns was the idea of keeping data at the farthest edge of the network until it’s needed.

I believe we’re heading for such a rapid proliferation of sensors and sensor data that we have to figure out new architectures that let us get the information we need without overburdening our networks. This feels like a tall order. Maybe Burns is working toward the solution.

Stacey on IoT | Internet of Things news and analysis

John Deere opens SF office to make its tractors smarter

John Deere’s office in San Francisco will eventually feature a simulator.

Farm equipment company John Deere is no stranger to the internet of things. It was connecting sensors and actuators on the farm twenty years ago. The next big thing in farming is what connected devices enable: precision agriculture. Precision Agriculture combines connected with devices with machine learning to have them make faster and more precise decisions, possibly without a farmer’s input.

To make precision ag the new reality, John Deere needs Silicon Valley skills. That’s why last week it opened an office in the SoMA area of San Francisco to connect with local talent.

Alex Purdy, head of John Deere Labs in San Francisco, says the hope is to find startups (or individuals) solving robotics, machine learning and artificial intelligence problems that may be of use in agriculture.

“We know the center of gravity for that talent pool is here and we want to organically build that,” he says.

Purdy says he has a list of 50 problems that he thinks machine learning, robotics and artificial intelligence can help solve ranked by how much impact they could have on farmers. For example, he’s looking for a way to deliver nutrients to plants at the right place and at the right time. Other efforts involve making better algorithms that can take advantage of the increasing number of sensors John Deere wants to add to its machinery.

The secondary goal of the office is to connect more deeply with Bay Area companies that already work with John Deere or its growers. For example, John Deere has a relationship with companies such as drone and satellite imagery company Mavericks; software providers FBN, FarmLogs and AgDNA and roughly 70 others.

John Deere is one of a long line of companies building Bay Area offices in hopes of attracting technical talent. Industrial heavyweights from GE to Ford have built up presences in Silicon Valley and San Francisco in the last decade as they realize the import of getting more technically savvy.

However, San Franciscans might find themselves surprised by how reliant farmers already are on tech and John Deere’s efforts to bring them to that point.

Stacey on IoT | Internet of Things news and analysis

Internet of Things News of the Week, May 26, 2017

Devices such as this WeMo Dimmer Switch will soon work with HomeKit thanks to a bridge.

Here’s a summary of interesting Internet of Things news from the past week. Get this summary in your inbox every Friday when you subscribe to the newsletter.

WeMo provides a bridge to HomeKit: Belkin, the maker of the WeMo-branded connected devices, announced it will release a bridge this fall that will allow its outlets, switches and other products to work with Apple’s HomeKit. Pricing wasn’t announced. As awkward as they are, I like bridges because when a company decides to build one it generally means they are all in on supporting multiple platforms. However, for a company like WeMo that only relies on Wi-Fi for connectivity to have to do this is a testament to both Apple’s market power and its utter arrogance. Props to WeMo for building a bridge instead of forcing consumers to buy all new HomeKit-certified devices.

New data on the smart home: Parks, an analyst firm, held its smart home focused event this week in California. It released a lot of data on the current state of the smart home such as the prediction that 442 million connected consumer devices will be sold in the U.S. in 2020. (Twice)

RedBull is connecting its coolers to the internet:  If you’re going to sell a best-served-cold drink to a crowd that’s literally addicted to your product, then you should at least make sure it’s kept chilled. That’s why RedBull is turning to the internet of things. The beverage company is connecting its coolers to monitor and ensure the temperature is right and it is also monitoring when the cooler door opens. RedBull is working with AT&T for connectivity and hosting the data as it arrives.

Why tools like AR and IIoT matter:  Adding intelligence to industrial settings clearly benefits operational efficiency, but this article gives a surprising stat about how it might affect workers. Yes, it will reduce the number of jobs, but it will also make the people who remain on those jobs more fulfilled. This article says that in a 10 hour shift a worker spends only 2.5 hours of that shift on productive work – work that adds value to the business. The rest of the time is used looking for information–meaning tools that can help the worker diagnose problems and solve them in the field could lead to a big increase in productivity. (IoT Central)

How to build a secure connected car: Car hacking is a big worry, even before we have autonomous vehicles driving us around. For those that want to understand the issues and see some thinking about how to protect vehicles, check out this report. (Cloud Security Alliance)

If you find notifications annoying today … just wait until your digital assistant is pinging you from your Amazon Echo or Google Home. The promise of such notifications is that they can share important information such a traffic accident that will affect your commute, even when your phone isn’t around. But it’s unlikely that you’ll want Alexa interrupting your dinner to tell you that someone just favorited your latest snapchat. (Fast Company)

Microsoft made a pair of decent AR glasses: Many companies are excited about the promise of augmented reality. Not for entertainment, but for its ability to overlay a lot of information for technicians. This can come in handy in industrial settings or even for doctors performing a surgery. The challenge so far has been finding a way to display information that leaves the user’s hands free. Microsoft’s new glasses demo is a step in the right direction. Yes, it looks like Warby Parker specs mated with a Goodyear tire, but the glasses look closer to normal eyewear than some of the other options out there. (Wareable)

Is it time for graph databases to shine? Last week I wrote about Google’s I/O event and how it signaled the shift to AI over software. Specifically, using AI to deliver context for customized services. This vision is compelling but, to make it real, I argued that we’ll need new distributed architectures that can remain flexible at scale. Looks like Microsoft is thinking hard about this problem, with a big focus on its Microsoft Graph as a way to connect everything from the nodes at the edge to each other and to the cloud. For insights on Microsoft’s efforts, check out this blog. (Wikibon)

Good news and bad news on wearables: For those wondering how useful their fitness trackers are, Stanford has some answers. The good news is that the heart rate tracking on these things is much better and accurate enough. The bad news is that the calorie estimates these things provide is essentially useless. This wouldn’t be a big deal expect that more and more insurers and corporate wellness programs are investing in fitness trackers to build incentives and bonuses that correspond with wellness. Wouldn’t it be nice if they were accurate? (Ashley Lab)

Microsoft and Johnson & Johnson execs call on government to regulate IoT Security: This happened late last week, but it’s worth noting simply because much of our thinking about IoT regulation stems from a knee-jerk dislike of government interference. But if done well, having a set of mandated rules could actually help the industry move forward and instill a sense of user confidence that their data is protected and the repercussions of sharing that data won’t come back to haunt them. (Data Center Knowledge)

Walt Mossberg is retiring: Uncle Walt was “the” reviewer for computers as far back as I can remember. His relevance may have waned as gadgets and gadget journalism exploded, but he’s a smart guy and his final column discussing the evolution of technology is a must read. (ReCode)

Verizon made a smart home hub: It has LTE as a backup, doubles as a Wi-Fi router and supports connected devices including Nest and Philips Hue. (Android Police)

Stacey on IoT | Internet of Things news and analysis