Google has recently introduced its enterprise Cloud Services Platform during the first day of its Cloud Next conference. The Cloud Next conference was held in the Moscone Center in San Francisco, CA. The tech giant has then announced two new products which according to the company will help the consumers to "develop and deploy intelligent connected devices at scale" on the second day of the event.
The first product which is the Edge TPU is Google's new IoT-focused hardware chip for the Edge platform. On the other hand, the second product which is the Cloud IoT Edge is a software stack which extends the Google Cloud's AI capability for the gateways and the connected devices. Google further says that collectively both the new products will enable users to "build and train ML (machine learning) models in the cloud, then run them on Cloud IoT Edge devices via the Edge TPU hardware accelerator".
The Edge TPU is the tech giant's proposed-built ASIC chip which according to the company is designed to run its TensorFlow Lite machine learning (ML) modes at the edge platform. Injong Rhee, Google's IoT VP, said that while designing the chip, the company was "hyperfocuses" in order to "optimize" the 'performance per watt' along with 'performance per dollar' within a small area.
"Edge TPUs are designed to complement our Cloud TPU offering, so you can accelerate ML training in the cloud, then have lightning-fast ML inference at the edge. Your sensors become more than data collectors-they make local, real-time, intelligent decisions".
Google has also introduced an Edge TPU development kit which includes a SOM (System on module) which collectively utilizes the Google's Edge TPU, an NXP CPU, Wi-FI, along with a Microchip secure element. Google further says that the feature will be available for the developers from this October.
The second project that the company has announced is the Cloud IoT Edge. The Cloud IoT Edge comprises two components i.e the Edge IoT core gateway functionalities and EDGE ML (Machine learning). The Edge ML is a runtime which is based on the lightweight open-source ML framework, TensorFlow Lite.
The feature can also run models on the edge devices which are based on the Android Things or Linux OS. The tech giant further says that it will extend Google Cloud's data processing and machine learning capabilities to the gateways, cameras and end devices. This will make the IoT applications more smart, secure and easily reliable.