We’ve been building large-scale industrial connected product systems since before any major cloud provider had services with “IoT” in their name, so it’s been an exciting few years as we’ve watched the industry grow and mature. Our (primarily) Fortune 1000 customers appreciate our open approach to IoT solution architecture, which provides the flexibility for running on their choice of cloud infrastructure and integrating with their existing enterprise systems and their choices of best of breed tools. We’ve connected their devices and data at the edge to their favorite cloud infrastructure and platform services for secure data ingest, device management, and other components as they’ve been released by the major providers. At the end of the day, the most important question is whether or not the solution produces value through Asset Management, Predictive Maintenance, and Yield Optimization.
The release of Cloud IoT Core Beta gave us the opportunity to add support to our customizable industrial reference application for Google Cloud Platform, the most recent of the public cloud providers to provide IoT services. We’re now able to offer customers a live IoT system on GCP, Amazon Web Services (AWS), or Microsoft Azure with initial dashboards and alerts, and event data made ready for integration with their enterprise systems in 30 days or less.
Following up on the announcement of our official support for Google Cloud Platform, we’d like to share our development experience and highlight the areas where GCP truly excels.
Data Ingress and Egress
Moving data from one system to another is a key portion of any IoT system architecture. GCP uses the mature and full-featured Cloud Pub/Sub as a core component for moving messages around between system components. The wide range of methods for getting data in and out of a GCP-based IoT System make it easy to integrate with existing systems and accept data from a variety of sources. The MQTT configuration that Cloud IoT Core has enabled with Cloud Pub/Sub is very developer friendly. The support for custom protocols at Cloud Pub/Sub as well as pull based requests and guaranteed delivery make it possible to unify brownfield legacy protocols with newer standards based protocols at the Cloud Pub/Sub layer. This sort of approach will be very powerful for anyone looking to retrofit an existing fleet of devices in addition to enabling their new product lines.
Permission & Trust Model
Security is of paramount importance in IoT Systems. One of the more challenging aspects of any IoT System is making sure the Permission Model is designed and applied correctly. The GCP model of permissions with sensible but overridable defaults values makes it straightforward to reason about the roles of a given service and user. This should help developers to avoid building systems that accidentally leave data accessible to unauthorized parties. IoT-specific features like Device Registries help developers organize and manage groups of device, further helping reduce the likelihood of leakage of data due to accidental misconfiguration of the system. The use of strong certificate protection provides both security and device identity validation, further enhancing trust in the system.
Device Management and Over-The-Air Updates
The ability to manage and update a remote device should be one of the first concerns addressed by any Thing maker while developing their IoT solution. GCP’s leveraging of Android Things is extremely powerful in that regard. Devices and gateways with sufficient resources to run Android Things gain a tremendous amount of capability by leveraging the Android community and associated mechanisms for remotely managing devices and updates. The Device Registry features help make management more straight-forward and reduce the likelihood for operator error when performing management functions.
World-Class Machine Learning
One of the primary ways of extracting business value from an IoT system is through actions taken based on insights generated by Machine Learning algorithms. Regardless of whether your goal is predictive maintenance, asset optimization, or yield optimization, the ability to generate business value is going to rely very heavily on the suite of Machine Learning tools that are available to the IoT system. GCP’s Machine Learning services are some of the best in the industry and give IoT products some of the best tools to generate business outcomes.
For more details about our experience with Google Cloud Platform and Cloud IoT Core, or to get started with GCP, AWS, or Azure using our customizable industrial reference application, contact us today.