As the IoT era keeps unfolding, tech molochs in Silicon Valley and everywhere else are brewing their solutions to solve the biggest problem so far: making sense of (so much) data.
A huge load of. Yes, because IoT implies a mountain of data so tall that the word “Big Data” may not be enough anymore to describe the zillions of terabytes in the jaw of cloud platforms nowadays. We all got used to the term “Big Data” when it meant data from the web – basically, from social networks. But this time, data is not about “content” – it doesn’t regard funny cats and memes, and the King, whether is content or not, is naked. Rather, it’s about the physical world around us, and the data that can get extracted from it. Sensors and gateways tracking you, and then engaging in conversation “about you” or, at least, for your own good. That’s the Internet of Things: a gargantuan Facebook for machines.
As expected, from Microsoft’s IoT Central to Samsung’s ARTIK Cloud, we’ve seen a swarm of big players jumping on the IoT wagon, and certainly Google isn’t staying at the window.
Google Cloud IoT Core – what is it and why it matters
Unveiled at Google I/O last Spring and released in public beta at the end of September, Google Cloud IoT Core is Google’s take on IoT. Mountain View has been fine-tuning the platform since its announcement in May, working shoulder to shoulder with a bunch of industries including, among the most prominent, Allwinner Technology, ARM, Intel, NXP and Realtek. The platform is already on and has served a plethora of complex use cases so far, from oil to gas, from utilities to transportation – even healthcare and ride-sharing.
According to Alphabet’s daughter company, the service is pretty straightforward. Apart from the usual promises of smoothless, scalable IoT connectivity, a strong point should be its immediacy: the user can ingest all his IoT data and connect them to “our state-of-the-art analytics services, including Google Cloud Pub/Sub, Google Cloud Dataflow, Google Cloud Bigtable, Google BigQuery, and Google Cloud Machine Learning Engine to gain actionable insights.”
TL;DR: this is not the umpteenth cloud platform, where you store your data to access them later. This is Google sharing its cutting-edge artificial intelligence tech with the everyday user. “Organizing the world’s information”, i.e. Google’s mission, doesn’t stop at stockpiling bits. Google will only thrive in the AI-dominated world of tomorrow if it will make sense of such bits. And this is where machine learning comes into the picture.
Data is the gold of the 21st century, but gold doesn’t mine itself. Machine learning mines this gold.
Now let’s move on to the core elements of the service.
Bring your own certificate
The problem Google is tackling here is accountability. Who can ensure that you actually own your device keys? Private beta users asked for this ability, and here it is. In addition to asymmetric key-based authentication for each device, users can import their device key signed by the device itself with the CA-certificate during the authentication process.
We talked about manufacturers, remember? This is where they enter the scene: manufacturers will be able to provision their devices offline in bulk with their CA-issued certificate, and then record the CA-certificates and the device public key on Google Cloud IoT Core.
HTTP connection for existing devices
The Internet as we know it wouldn’t exist without the Hypertext Transfer Protocol – “HTTP” for friends. So it shouldn’t raise any eyebrow that Google is trying to leverage HTTP’s ubiquitous popularity for connecting existing IoT devices and gateways to its service. In other words, everybody knows HTTP – so it makes sense to complement the standard MQTT protocol with its older brother. According to the company, the feature will ensure a more secure connection and simplify the process of feeding data to the platform at scale.
Logical device representation
Even in the best of families, sometimes the Internet just doesn’t work. But if this may drive your son crazy, that’s a whole different ballpark if your industry serves millions of actual people worldwide. Logical device representation is a solution to this problem. Thanks to the feature, the user can retrieve the last state and properties of an IoT device even if such device is not connected. And it doesn’t end here: not only users, but most importantly APIs may retrieve them.
The logical device representation has enabled many enterprises to build new IoT solutions, among which the company mentions energy management, transportation and logistics.
Pricing
Mountain View’s data plan follows the freemium model we have come to expect of it. Despite that, it’s worth noting that for the company ensuring a free tier is important also in the IoT domain.
Case History: Smart Parking Limited
Smart Parking Limited is a world leader in the design, development and management of parking technology. But the way to solve the parking problem is different nowadays than ten years before. Smart Parking Limited, aware of the new trends, has been transforming the parking experience over the last ten years, ensuring increasing revenues for a huge pile of customers. As you’ve probably noticed, this is not just about smart parking: Smart Parking Limited is becoming a smart city company.
But why are we talking about Smart Parking Limited?
The enterprise has been part of Google Cloud IoT Core’s public beta and built its new platform on top of it. Through the platform, Smart Parking Limited enables seamless integration and connectivity between distributed devices within a city. What’s more, Smart Parking leverages Mountain View’s machine learning algorithms to run predictive analytics on the whole city.
Conclusions
Google is maybe the fastest mover in this space, given that some of its competitive services, such as Microsoft IoT Central, aren’t available yet. On the contrary, Google’s platform is up and running, and from the news we have is not only about making things smart. Because you know, that’s what IoT has been about.
For years.
But Google is doing something different: through machine learning it wants to put your mind at ease and help you make more informative decisions.
Basically, Google wants to make you smart.