Tradeoffs - IoT Lens

Tradeoffs

IoT solutions drive rich analytics capabilities across vast areas of crucial enterprise functions, such as operations, customer care, finance, sales, and marketing. At the same time, they can be used as efficient egress points for edge gateways. Careful consideration must be given to architecting highly efficient IoT implementations where data and analytics are pushed to the cloud by devices, and where machine learning algorithms are pulled down on the device gateways from the cloud.

Individual devices will be constrained by the throughput supported over a given network. The frequency with which data is exchanged must be balanced with the transport layer and the ability of the device to optionally store, aggregate, and then send data to the cloud. Send data from devices to the cloud at timing intervals that align to the time required by backend applications to process and take action on the data. For example, if you need to see data at a one-second increment, your device must send data at a more frequent time interval than one second. Conversely, if your application only reads data at an hourly rate, you can make a trade-off in performance by aggregating data points at the edge and sending the data every half hour.

IOTPERF 9. How are you ensuring that data from your IoT devices is ready to be consumed by business and operational systems?
IOTPERF 10. How frequently is data transmitted from devices to your IoT application?

The speed with which enterprise applications, business, and operations need to gain visibility into IoT telemetry data determines the most efficient point to process IoT data. In network constrained environments where the hardware is not limited, use edge solutions such as AWS IoT Greengrass to operate and process data offline from the cloud. In cases where both the network and hardware are constrained, look for opportunities to compress message payloads by using binary formatting and grouping similar messages together into a single request.

For visualizations, Amazon Kinesis Data Analytics enables you to quickly author SQL code that continuously reads, processes, and stores data in near-real-time. Using standard SQL queries on the streaming data allows you to construct applications that transform and provide insights into your data. With Kinesis Data Analytics, you can expose IoT data for streaming analytics.