Amazon AWS IoT

AWS IoT is the managed cloud platform of IoT.

2019: The joint solution Nvidia and Amazon based on Jetson and AWS IoT Greengrass

The Nvidia company announced on March 18, 2019 cooperation with Amazon company according to the NVIDIA Jetson platform to allocate with artificial intelligence and deep learning millions of attached devices. This joint solution allows to create, train and optimize models on AWS and then to develop them on end devices based on Jetson using AWS IoT Greengrass.

The NVIDIA Jetson platform offers opportunities to AI in the end device with high-performance and economic computing. It received application in independent machines and smart cameras in such industries as retail, production, agriculture and others, noted in Nvidia.

AWS IoT Greengrass accurately transfers AWS to end devices, including inferens machine learning that it was possible to apply it locally to the generated data, at the same time using a cloud for management, analytics and reliable storage. Devices based on Jetson execute inferens locally to receive result almost instantly using AWS IoT Greengrass. Then data go back to services of machine learning, such as Amazon SageMaker, for increase in accuracy of model.

Jetson is delivered with a full-fledged set of software tools and SDK, including NVIDIA JetPack. Also the different frameworks, such as MXNet, Caffe, TensorFlow and PyTorch allowing developers to use these algorithms for fast deployment of applications of the real world are supported, claim in Nvidia.

"Jetson is the high-performance computer constructed on the same architecture and the unified software as powerful supercomputers. We offer small Jetson Nano in the help to Jetson AGX Xavier that it was possible to scale small IoT of the device to powerful IoT of gateways". Deepu Talla, the vice president and the director of independent machines in NVIDIA

As noted in Nvidia, Jetson and AWS IoT Greengrass allow to develop the models of machine learning optimized for IoT of devices with ease. Clients can save traffic and expenses, executing inferens practically in real time directly on the Jetson device, but not sending data to a cloud.

According to the statement of Nvidia, Jetson allocates with artificial intelligence of the application, earlier seeming unreal. In camera agriculture based on Jetson with the started AWS IoT Greengrass can find weeds practically in real time, remove earlier indefinite weeds, load anomalies into a cloud and quickly retrain and develop model. This combination of technologies does agriculture more effective and allows to solve a problem of prosperity of food for increasing population. Automatic optical check allows to define quickly defects in products directly on production to avoid delays in operation of the pipeline. It increases productivity, minimizes losses and, finally, improves overall effectiveness of work. In retail of the device based on Jetson can check warehouses, keep track of behavior of buyers and take and process data on the end device, applying AWS IoT Greengrass to deployment of the trained neuronets. It allows to increase efficiency and in retail.

2015: Announcement of the platform

On October 14, 2015 the Amazon company announced service which will be called AWS IoT.

AWS IoT is the managed cloud platform which, by description developers, "will allow attached devices easily and to interact safely with cloud applicaions and other devices"^[1].

Representation of AWS IoT (2015)

For October 15, 2015 the IoT platform works in the beta, will be capable to support billions of devices and to process trillions of messages sent from devices to AWS and to other devices. It should be integrated with other services of provider:

• Amazon Lambda,
• Kinesis,
• S3,
• Machine Learning
• DynamoDB.

It will allow to create the IoT applications, to manage infrastructure and to analyze data. AWS IoT will also provide SDK to third-party application developers.

The cost of IoT-service, by analogy with other services of Amazon Web Services, will depend on number of the messages (data units on 512 bytes) sent to AWS and the messages sent by the platform to devices or applications without minimum payments. As a bonus for users of an ecosystem of AWS cost does not join the messages sent to Amazon services, in particular S3, DynamoDB, Lambda, Kinesis, SNS and SQS. The payment, depending on the region, will make from $5 to $8 for 1 million messages.

Notes