When leading-edge organizations first started to scale their IoT initiatives, some developed and deployed their own code to manage the endpoint devices and generate data as it moved from device to cloud.
That software work often took months to complete. Now, organizations can use IoT platforms from Microsoft, AWS, Google and other vendors instead of building their own, giving enterprise IoT engineers the ability to stand up the technology in mere days.
These platforms provide secure, resilient and flexible connections to endpoint devices and deliver a range of management capabilities that simplify and streamline IoT deployments.
"It's an aggregation point for unattended devices to connect and be managed," explained Brad Cole, director of engineering at Digi International, an IoT tech organization. "A [platform] provides a common way of communicating with all the IoT devices and a common way for the devices to send data that can then be distributed through the ecosystem."
Why are they needed?
The difficulty and challenges of managing a growing IoT environment as organizations move from small-scale pilots to full-fledged deployments, where hundreds and even thousands of endpoint devices, the data they generate, the analyses of that data and integration with the cloud as well as other systems all must be managed and maintained.
IoT engineers and developers could build their own systems to do that work, just as early adopters of IoT did, or they could use APIs to manage the devices as well as extract, direct and analyze the data generated by the devices.
These platforms provide the various capabilities needed to manage an IoT ecosystem at scale, freeing developers from having to build management functionality themselves.
"But that's very complex," said Luke Durcan, director of EcoStruxure at Schneider Electric. Homegrown solutions are expensive to create and maintain and often fail to offer key capabilities such as security. That's where IoT platforms step in, he said.
These platforms provide the various capabilities needed to manage an IoT ecosystem at scale, freeing developers from having to build management functionality themselves. Instead, they can focus on creating the logic and any other unique code required for their organizational needs.
"[An IoT platform] is necessary as you start scaling, because as you start connecting more, you have to invest in an ongoing way to manage these assets. You have to figure out what connections you need, how to scale the platform, whether it's wireless or wireline. You have to think about updating these connected assets and troubleshooting," said Michele Pelino, a principal analyst at Forrester Research and a member of the firm's infrastructure and operations research team. "That complexity is why you need a [platform]."
How do IoT platforms work?
According to Pelino, organizations use IoT platforms -- or, as Forrester labels them, industrial internet of things software platforms -- to handle one or more of five major functions: connecting to the devices, securing them, managing them, supporting analysis of the data, and enabling developers to create code and applications that interact with the IoT system and integrate with broader enterprise systems.
Vendors typically deliver these platforms as a service. They offer either a standard package of functions for their enterprise customers to implement within their IoT deployments or a menu of options for customers to pick and choose which capabilities they want. Experts said they're essential components of the IoT ecosystem.
"Companies are bringing all these different IoT devices together to manage and provision, and they want everything to communicate with each other," said Dilip Sarangan, senior director of research at market research company Frost & Sullivan. "An IoT platform manages the devices across geographies and multiple systems. It's multiple layers of software that makes it seem like it's one level."
IoT platform options
Many of the big enterprise technology companies, as well as the public cloud providers, offer IoT platforms.
The Azure IoT Hub from Microsoft, the AWS IoT Core and the Cloud IoT Core from Google are among the dominant offerings in the market. SAP, Oracle and Siemens also have IoT platforms.
There are, in fact, hundreds of vendors offering IoT platforms that have one or more of the five core capabilities identified by Forrester.
"[IoT platforms] leverage cloud-based computing resources to enable interaction with the end IoT device and end-user interaction," said Bob Reny, EMEA CTO and principal engineer at the tech organization Forescout Technologies. "[They] are all managed by a stack of vendors via the cloud ecosystem. This includes storage, systems, security and application management. All services come at a price, but these cloud-based resources are cost-effective, often reusable for other peer services and can accelerate any IoT delivery to consumers."
Choosing the right IoT platform
IoT platform vendors offer differing ranges of capabilities.
"The big cloud ones give you flexibility; smaller niche players tend to be more focused," Durcan said.
Many organizations decide to use the IoT platform made by their public cloud provider to simplify and streamline the connections between endpoint devices, the platform and the cloud.
Although opting for an IoT platform from one's own cloud provider isn't necessary, Sarangan said organizations can avoid additional cost and complexity that often comes with using one vendor for the IoT platform and another for cloud services.
Organizations can also use more than one IoT platform. Experts said this scenario is more likely in organizations using a niche platform that provides functionality needed for a particular use but that doesn't provide the full scope of services required by the enterprise IoT deployment.
Enterprise leaders must evaluate their IoT initiatives and purpose and look for a platform that can best deliver the capabilities needed to meet its objectives. They must then consider which vendor options fit best within their IoT deployment.
"You need to know what your requirements are or you might have complexity you don't need," Pelino said.
