Analysis and Up-gradation of Batch Fusion Development in RAF

Abstract

There has been a massive growth in the retail software market worldwide. This is mainly due to the expansion of the retail software in the emerging economy. Also the enterprise retailers are re-architecting their systems to adapt themselves to the changing technological requirement. With the fusion applications migration to SaaS, the batch execution requires a scheduler to automate the batch running. It also requires auto monitoring capabilities to ease out the batch execution operations via AI techniques. Monitoring module utilizes the scheduling capabilities of Batch scheduler to drive the batch execution cycles within and also between retail applications. The key engines to automate the batch running are POM and JOS, where one can run batch jobs and other can monitor and accordingly generate notifications for updating the status for the batches to the customers. In the process of automating batch running, three modules namely Job admin, Process ow and Batch scheduler are used. Each of the three modules are provided with artifacts as seed input which is generated by AG. AG is a semi-automated tool which takes XSLM sheets as an input and generates artifact to be fed to each of the three modules respectively. Since all the applications are deployed independently in distributed systems, there arise a need to use ReST calls for transferring data to establish inter-application communication. The ReSTful implementation of Retail Platform Services is done by the creation of a separate service layer followed by the data access object layer that has a direct interaction with the database. As ReST calls are expensive, some optimizations are required. From the performance perspective, the Retail Platform Services are deployed using the weblogic server.