Designing an architecture for optimal edge deployments

While there are a number of key factors that influence the edge architecture design, as was highlighted in the previous section, there is also a critical capability necessary to enable these factors, namely the ability to build, deploy, and manage the device software at the edge. Additionally, we also need the ability to manage the application, in essence, the ML model deployed to run on the edge devices. Consequently, AWS provides both of these management capabilities using a dedicated device management service called AWS IoT Greengrass (https://aws.amazon.com/greengrass/), as well as the ML model management capability built into Amazon SageMaker called Amazon SageMaker Edge (https://aws.amazon.com/sagemaker/edge). AWS IoT Greengrass is a service provided by AWS to deploy software to remote devices at scale without firmware updates. Figure 8.1 shows an example of an architecture that leverages both Greengrass and SageMaker to...