Connecting to Live Data in SAP HANA - Overview
Live connectivity support in Smart Predict allows you to train and apply a predictive model using business data that stays in your on-premise SAP HANA system. The predictions are also stored in your on-premise SAP HANA system. Up until now the data had to be uploaded, or acquired in SAP Analytics Cloud (SAC), before being processed within the SAC platform. Live connectivity support means that your business data stays where it is, removing the need for physical data movement.
The following diagram shows how the different components of the live data connectivity fit in
with each other:
Restriction
You can use
Smart Predict on SAP HANA on-premise only. Cloud
deployments of SAP HANA are not supported currently. Using Smart Predict on live SAP HANA Data is based on the
following principles:
- Your business data stays within the on-premise SAP HANA system when a predictive model is trained or applyed from SAP Analytics Cloud.
- The execution of the predictive algorithms is delegated to the on-premise SAP HANA system using the SAP Automated Predictive Library (APL).
- The data used to train and apply your predictive model, the predictive models, the predictive model metadata (such as the performance indicators and the influencer contributions) and the predictions are stored in the on-premise SAP HANA system.
- Once trained on a specific system, a predictive model can be applied to data sitting on the same system. Predictions resulting from applying predictive models are also stored in tables in the same system used for training.