SAP SE
PREDICTING DOWNTIMES FOR SOFTWARE SYSTEM UPGRADES

Abstract:

Techniques for predicting the downtime of a software system due to an upgrade of the system are provided. In one set of embodiments, a computer system can receive training data comprising (1) measured downtimes, (2) upgrade object information, (3) upgrade configuration information, and (4) system hardware information for past upgrades of the software system across different operating environments. Using this training data, the computer system can train a machine learning (ML) model to predict (1) based on (2), (3), and (4). The computer system can then receive a query comprising upgrade object information for a future upgrade U, upgrade configuration information for upgrade U, and system hardware information for an operating environment E. Based on these inputs, the computer system can use the ML model to generate a downtime value indicating the predicted downtime that will be experienced by the software system when upgrade U is applied to the system in operating environment E.