Cloud services are an essential slice in the architectural pie of IOT. Data gathered from things (devices) need to be processed and distributed to enterprise applications for human consumption. Cloud with its IOT specific platform stack facilitates just this.
Top 3 cloud providers have their IOT stack lined-up. Here’s a brief look at each of them.
- IOT edge for device side processing (real time analytics with ML capabilities).
- Android things – OS for devices.
- Hardware Edge TPU (Tensor Processing Unit) for device side AI.
- Cloud IOT core for device management, bi-directional communication.
- Cloud pub/sub for distributed message processing.
- Cloud dataflow for stream processing.
- Big data stack such as cloud function, ML, big table, big query and visualization tools such as insights.
- Free RTOS (Real Time Operating System) for devices.
- IOT dash button – Programmable IOT thing.
- AWS Green grass – Bringing cloud capabilities such as ML, lambda into the device side. Lets setup a group in the field for devices to connect and work in sync.
- IOT core as communication gateway, device shadowing, device provisioning, etc.
- Big data stack involving message processing (Kinesis), storage, lambda, etc.
Link – AWS IOT ebook
Microsoft’s Azure provides
- IOT edge device for supporting device side processing.
- IOT hub as gateway for communication, monitoring and device management.
- Supports the typical message processing, stream computing, big data analytics and event delivery.
Across the cloud providers, we see a pattern as below.
- Capabilities being added to the things (device) side for making them smart supporting real-time intelligence and better sync with cloud – OS, SDK, edge computing, edge ML, etc.
- Device management capabilities such as device shadow, provisioning, pushing updates, security, etc from cloud side.
- Gateway to communicate with the devices in the field.
- Support for stream processing and typical big data stack including analytics and visualization at the cloud side.