Cloud Services for IoT Buyer’s Guide
Summary
To help you decide which of the Big Three cloud services providers – Google, AWS or Azure – is best for your IoT project, we researched performance, pricing and other relevant information and compiled it into this handy Buyer’s Guide.
Deploying and managing IoT devices is hard. Public cloud vendors are trying to make it easier by offering cloud-based services that simplify the tasks of keeping track of all of the devices in a company’s IoT network, uploading and downloading data from them and processing that data.
Each of the Big Three clouds now include an IoT service for this purpose. On AWS, it’s called IoT Core. Google Cloud calls its service Cloud IoT Core. And Azure offers IoT Hub. Each cloud also provides some complementary IoT-related services.
At a high level, these cloud IoT platforms provide the same functionality. They let organizations create registries of IoT devices to help manage them and add them to an IoT network. The services also make it possible to collect data from IoT devices, process the data in the cloud and take action based on the results.
But when you dive into the details, it becomes clear that each cloud provider’s IoT service has its own quirks. To help businesses decide which cloud IoT service is best for their needs, we’ve prepared a Buyer’s Guide that summarizes performance, pricing and other relevant information for each of the Big Three public cloud’s IoT services.
Here’s an overview of the trends you’ll find in the guide.
Scalability
Each cloud’s IoT service is designed to support large-scale IoT networks that contain a million or more devices. However, there are some feature differences that could impact the relative scalability of one IoT platform compared to another.
One factor is the number of devices that organizations can manage through each cloud IoT platform. AWS IoT Core and Google Cloud IoT Core impose no limit on this figure. Azure IoT Hub imposes a limit of 1 million devices, although, according to its documentation, more devices can be added upon special request.
Also relevant for scalability is the number of concurrent connections that a cloud IoT platform can support. This factor is important because, although organizations can have very large numbers of devices registered with a cloud IoT service, the number of connections they can have active at a given time is smaller. Thus, your ability to use all of your devices at once is limited by the number of concurrent connections that your cloud supports.
AWS supports up to 500,000 concurrent connections without any throttling limits. Google Cloud, meanwhile, supports only 60,000 concurrent connections.
On this front, there is significant variation between the vendors. Azure’s solution is theoretically the most scalable, because it supports up to 1 million concurrent connections; however, the caveat is that Azure imposes what it calls a “device connection throttle” to limit connectivity speeds when the number of concurrent connections becomes very high. AWS supports up to 500,000 concurrent connections without any throttling limits. Google Cloud, meanwhile, supports only 60,000 concurrent connections.
Network Performance
Another critical consideration for a business managing an IoT network is the network performance rates of the cloud that manages that network. Slower performance by the cloud provider’s network will negatively impact a business’ ability to move data between its IoT devices and the cloud quickly.
AWS and Google both make throughput guarantees of 512 KB per second per client for data uploads and downloads. Azure doesn’t provide a throughput guarantee, however, which makes its performance less certain.
On the other hand, Azure IoT Hub does support one communication protocol for exchanging data -- AMQP -- that the other two providers don’t. (All three providers support two other common protocols, MQTT and HTTP/HTTPS.) Although no one protocol is necessarily faster than others, having more communications options gives developers more opportunities to optimize network performance and latency. In this respect, Azure’s IoT platform has a potential performance advantage.
Availability and Uptime
Perhaps surprisingly, we couldn’t identify any specific availability or uptime differences between the various cloud IoT services. Each vendor guarantees uptime of 99.9% for its cloud IoT service.
The overall cloud uptime histories of each vendor’s infrastructure do vary, as we’ve noted in the grid. However, none of the vendors publishes uptime history for its IoT service specifically, so hard-and-fast conclusions about availability performance for IoT services in particular are difficult to draw.
Pricing
As is true of most types of cloud services, comparing prices between cloud IoT platforms from different vendors is tricky. Each vendor structures its IoT service pricing differently in terms of what it charges for: AWS bills for connectivity time as well as messages sent, while the other two clouds charge only for messages sent. To make matters more complicated, the default message size assumed by each vendor’s pricing schema is different. And the pricing tiers from each vendor also vary.
To help summarize pricing differences, we’ve included pricing details for the lowest-cost paid IoT service from each of the cloud vendors. We’ve also noted nuances in message size as they relate to pricing in order to enable something that approaches apples-to-apples price comparisons. In general, AWS IoT Core offers the lowest per-message pricing, but (as noted above) it adds connectivity fees to this figure. Google Cloud IoT Core’s per-message costs are the lowest in general and are not accompanied by connectivity fees.
Overall Cloud IoT Trends
At the risk of overgeneralizing, here are some overarching IoT cloud trends that are clear from the data we’ve collected:
- Google Cloud IoT Core probably offers the lowest prices overall, but its scalability is more limited.
- Azure IoT Hub offers the best theoretical scalability because it supports up to 1 million concurrent connections, but throttling might limit the performance of the service for large-scale IoT networks.
- AWS IoT Core offers a high degree of scalability and performance, and its pricing per message is attractive; however, the fact that it charges for connection time as well as per message makes costs somewhat harder to predict.
Christopher Tozzi,
IoT World Today
Christopher Tozzi is a technology analyst with subject matter expertise in cloud computing, application development, open source software, virtualization, containers and more. He also lectures at a major university in the Albany, New York, area. His book, “For Fun and Profit: A History of the Free and Open Source Software Revolution,” was published by MIT Press.