AWS threatens IIoT partner ecosystem
This past week, AWS has kicked off their reInvent conference; this year a 3 week virtual marathon. In the first week alone, AWS has made a remarkable 65 announcements of which 4 appeared to be relevant to IoT and Edge computing. These announcements make it clear that AWS is following a vertical integration strategy that leaves very little room for partners.
Let’s first look at the 4 announcements relevant to IoT:
- Amazon Monitron is billed as a end-to-end equipment monitoring system that includes sensors and gateways that can be installed in factories. For only $715 you can purchase a starter kit from Amazon. Amazon is positioning Monitron as a predictive maintenance solution for factories. The Amazon hardware sends the sensor data to the AWS Cloud for machine learning analysis and reporting through a mobile app.
- AWS Panorama is another piece of hardware for machine learning of computer vision. Again the use case is in the factory to evaluate quality of the manufacturing process. AWS Panorama appears to be able to work with existing cameras on the factory floor.
- Amazon Lookout for Vision appears to be similar to AWS Panorama but is a ML service running on the AWS cloud. Again the use case for this service is quality control for production processes. You send a set of images to the AWS cloud for ML processing.
- Amazon Lookout for Equipment appears to be similar to AWS Monitron but once again a ML service on the AWS cloud.
In summary, AWS announced hardware and machine learning services that target some of the main use cases for industrial IoT: predictive maintenance and quality control through computer vision.
Up until now, AWS was growing an extensive partner network of hardware vendors and edge computing specialist that are targeting these use cases. Many of these vendors were happy to use AWS and AWS IoT as their cloud service and add value through specialized hardware and specialize machine learning models. Now that AWS has entered this market directly, the partner ecosystem around AWS IoT will likely be disrupted. AWS wants to own the entire value chain so they will work to displace many of these companies.
For many companies, especially small factory operations, a complete end to end solution will make sense. Many of these small factories do not have technologist on staff. They rely upon consultants to handle the technology requirements of operating their factories. These consultants will see AWS as being the cheapest and easiest solution to implement. Larger companies will have more advanced requiremetns and technology staff that will select the best solution available. The partner network of hardware vendors and ML vendors will continue to target the higher end but over time AWS will move up the value chaing to compete with these customers too. Eventually this will lead to industry consolidation over time.
AWS is proving to be a competitor through-out the entire IIoT value chain. In the long-term, I expect to see IIoT vendors looking for other IoT cloud solutions that they can offer to their customers to replace AWS IoT. It won’t make sense for these vendors to introduce their customers to AWS, if in the future AWS will attempt to replace their offering.
I see this as an opportunity for MQTT and Sparkplug to become even more important in the IIoT domain. A standard protocol and a standard payload definition allows for an ecosystem of equals amongst vendors to offer customers the best of breed solution that match their requirements. This is the vision we have at HiveMQ and are very excited to see the momentuem we are seeing with customers that want to embrace these open standards. Getting locked into an AWS end-to-end solution is not going to be a good long-term strategy for anyone, except AWS.