It seems like everyone wants to get into the IoT business these days. The top three cloud vendors are no different, as they continue to use their market advantages to promote their offerings.
And despite the similarities in the cloud providers' IoT services, there are substantial differences. For example, AWS, Google Cloud and Microsoft Azure differ in how they package IoT tools and work with partners. Furthermore, in some cases, an enterprise may find a newer or smaller competitor outside of the big three better fits its needs.
In this IoT cloud platform comparison, we'll provide an overview of what's available from the major providers and how the vendors' IoT platforms differ. We'll also help you weigh your options when it comes to cloud versus industrial IoT.
AWS vs. Azure and Google
The AWS platform offers good fundamentals that provide the building blocks for scalable, reliable and secure IoT applications, said Tolga Tarhan, CTO at Onica, an AWS consulting partner. Developers can let the AWS IoT platform handle the heavy lifting and focus on making their application or feature unique.
As recently as 2018, some experts believed Google Cloud Platform (GCP) trailed AWS and Azure in IoT capability. That year, Forrester Research produced a report on industrial IoT and cloud that examined a dozen or so providers, including AWS and Azure -- but not Google.
"We didn't feel that Google had a sufficiently rich set of capabilities in production, nor a sufficiently broad and deep set of reference deployments with paying customers," said Paul Miller, Forrester analyst. However, he added, Google has come a long way since then.
Nowadays, in terms of an IoT cloud platform comparison, there's significant overlap among AWS, Google and Azure, according to Jonathan Dexter, engineering practice director at Nerdery, a digital business consultancy. For instance, Azure IoT Hub, AWS IoT Core and Google Cloud IoT Core all offer cloud and edge capabilities, as well as device management. The core components are often the same, though they're packaged differently, Dexter said.
For example, Azure offers IoT Central, an environment with a message queue, device management and event triggers within one platform. Similar features are available in GCP or AWS, but developers working on those platforms would need to build those features out themselves.
Likewise, each cloud IoT environment offers its own analytics tools, both streaming and in-place, along with visualization technologies. While Google offers Cloud Machine Learning and Data Studio, Azure offers Power BI and Streaming Analytics.
IoT ease of use and functionality
Organizations will largely find it easier to use the cloud they're already familiar with, though there are differences in platform functionality. For example, Azure's IoT Central provides simplicity, but it's constrained in the problems it can solve, Dexter said.
"Each provider requires a different collection of edge, messaging, streaming analytics or business intelligence functionality, and making a from-the-ground-up IoT platform requires knowledge of software development," he said.
Azure has placed significant focus on packaging IoT platforms with a comprehensive set of partnerships. AWS IoT is a close second, but requires some aptitude with the AWS platform.
Google is an up-and-coming IoT contender, Dexter said. It can use its machine learning expertise, featured in BigQuery and TensorFlow, to power and differentiate its IoT capabilities.
"TensorFlow and Azure IoT Edge both aim to target the requirement for edge computing with focuses on things like image recognition or contact signaling," Dexter said. AWS IoT Greengrass has a similar deployment model with [email protected], but with less direct platform focus on machine learning and analytics, he added.
IoT cloud platform comparison: Beyond the big three
Industrial players such as ABB, Bosch, General Electric, Hitachi, Schneider and Siemens have recognized they cannot compete directly with the hyperscale cloud providers to build a global network of data centers, Forrester's Miller said. Instead, they've opted to run their industrial IoT platforms on top of the major public clouds.
Initially, many of the industrial companies used these clouds for little more than hosting. But now, they're incorporating more of the public cloud IoT capabilities into their services as they seek to focus on unique capabilities they can offer, Miller said.
For example, these industrial vendors have expertise in how to connect and maintain physical IoT devices. They may also have private cloud or edge versions of their IoT software for when the public cloud isn't quite up to the job due to intermittent connectivity, bandwidth limitations or data territoriality concerns.
Cloud providers still have an advantage in analytics and machine learning. IoT platforms can't just be about connecting and managing IoT devices. They need to support workflows that extract insights from the data. Cloud providers often have a richer set of analytics and machine learning capabilities than the industrial companies.