Synopsis
For those familiar with SAS Analytics for IoT, you probably know that the current ingestion process of time-series data is driven by a SAS macro called afi_dataload, which is shipped with the SAS for AIoT software. As its input, the macro takes a simple parameter file that contains a series of options that point to a variety of files identifying measures, devices, events, and (optionally) hierarchies that describe relationships between entities belonging to each of the aforementioned categories.
The attached document describes the improvements that the enhanced version of this ingestion process brings to the table. The goal is to make the solution more flexible by, for instance, extending support to every data source accessible through SAS Viya, either via connectors, or through SAS Event Stream Processing (ESP) adapters. Other features include the ability to take backups of the SAS AIoT Data Mart tables as well to filter data directly at the source when the data sample required to train a particular model is narrow in scope.
To get more details about these and many other new features we invite you to look at the attached document. A ZIP file is also attached containing the macro code, a file showing sample invocations covering different scenarios, as well as the Job Definition and Job Form files that are needed in case you want to invoke the macro directly from your Web Browser.
Finally, we would like to kindly ask for your feedback as your opinion is the main driving factor in the development of tool kits like this one. Let us know what you think of the existing features, and what others you would like to see added to further improve the functionality of the solution.
Thanks,
Mauro Cazzari
Sr Principal Solutions Advisor
Pre-Sales Support Internet of Things (IoT)
