While the average quality of IoT platforms continues to increase, finding platforms with unique value-creating capabilities and independently benchmarked scalability will be difficult in 2020.
2019 has been an amazing year for those of us who spend our business lives using IoT platforms. While prior years have been replete with big promises and little to show, 2019 has been a year where everyone is running at full gallop to deliver meaningful implementations. Designing IoT system architectures, scaling proof-of-concepts (POCs) to pilots, and working on systems integrations have become everyday occurrences.
MachNation has noticed some interesting changes in the IoT platform world in 2019 as vendors have polished their solutions. This blog discusses what we have seen in the past year and implications for 2020.
Prediction 1: Most IoT platforms will be good enough for the average IoT deployment.
No longer are we in a world where it is difficult to find an acceptable IoT platform. The average IoT platform will provide an enterprise with sufficient tools to implement a small-scale IoT deployment, or at least successfully complete a simple POC with limited consternation.
Overall, the average IoT platform in 2019 has a set of table-stakes capabilities including IoT data management, access control, condition monitoring, visualization, support for the MQTT protocol, support for a minimal number of IoT devices, simple developer documentation (or at least developer docs that are correct), and a way to effectively automate the bulk management of devices.
If an enterprise just needs an IoT platform that is “good enough”, then finding one should not be too difficult.
That leads us to prediction 2.
Prediction 2: It will require a lot more work to find best-in-class usability and exceptional capabilities for edge, device management, and event processing.
While the quality of the average IoT platform has increased from 2018 to 2019, it is becoming increasingly difficult to identify truly unique and valuable differentiators between IoT platforms. This is a pretty typical phenomenon in maturing technology sectors.
It is important that an enterprise identifies points of differentiation in an IoT platform, especially if an enterprise’s requirements will include:
- An exceptional user interface, including a well-designed front-end allowing OT users and other platform personas – especially those at mid-market enterprises – to engage with the platform
- IoT edge capabilities, including the need to easily push an application or advanced analytics capabilities from the cloud to the edge
- IoT device management capabilities, including the need to support myriad device protocols and offer a productized way to deploy devices at scale
- IoT event processing, including the need to have developers create rules to easily integrate business and device-data logic into the management of IoT assets.
It is also fair to say that developer documentation can make-or-break an IoT platform. We are constantly surprised to learn about poor documentation, even in the largest hyperscale cloud vendor’s IoT docs.
Developer docs should do more than just answer the question, “How do I complete this task?” They should help the developer determine the best way to complete a task and describe why one approach is preferred to another. The documentation should have tutorials, videos, or other material that help streamline the development effort and save developer time.
Prediction 3: Platform scalability and performance matter, though vendors do not have a way to independently provide this information.
Platforms are not created equally when it comes to scalability and performance of the underlying microservices. If your IoT platform vendor has used words like “unlimited scalability” or “five nines of service quality”, you can be assured the vendor is pulling your leg.
IoT platform vendors have not stress tested their platforms in any meaningful way. They have no idea of message failure rates when their platforms ingest messages from 10,000 or 10 million devices. They do not know data latency rates when their event processing engine handles 10 event rules for a system of 100,000 connected devices. They have not tried pushing over-the-air firmware updates to 5,000 devices simultaneous while measuring failures and latencies.
Ask your chosen IoT platform vendor for independent data verifying the platform’s scalability. Then listen for the silence.
As average IoT platforms become increasingly homogenized, performance and scalability will matter more in 2020 and beyond. Enterprises selecting an IoT platform should insist on having verified, independent scalability metrics from their chosen vendor. And enterprises that have already deployed IoT solutions should be testing their implementations to ensure their chosen platforms do not fail at scale
Conclusion
2020 will be another year where enterprises are focused on implementations and platform selection. While overall quality of platforms across the ecosystem has increased in 2019, finding exceptional IoT platform capabilities is becoming more difficult. And finding IoT platforms that can truly scale is like finding a thoroughbred champion among a pack of barnyard horses.