Batch Processing Error Handling via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional business object processing systems face challenges in handling batch operations, where errors in one business object instance can lead to the entire batch job being aborted, requiring resubmission, rather than allowing successful operations to proceed independently.
Innovation Solution
A workflow that partitions batch processing into stages with error checking and filtering, allowing successful business object instances to continue processing while failed instances are cleaned up and retried, utilizing a service manager, batch coordinator, and service provider to manage transactions and errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional batch processing is used where errors in one business object instance abort the entire batch job, then data integrity is maintained, but productivity is reduced due to full job resubmission
Solution Approach 1:
The batch job is segmented into individual business object instances that can be processed independently. Each instance is evaluated separately, allowing successful instances to be saved while failed instances are identified and removed, enabling partial batch success rather than complete abort
Solution Approach 2:
Instead of requiring complete batch success or complete resubmission, the system performs partial action by saving only the successful business object instances. This allows the batch processing to proceed with the portion of work that succeeded, eliminating the need for full job resubmission
2Stability of the object's composition
If the entire batch job is resubmitted after any error occurs, then consistency is maintained, but loss of time increases due to reprocessing successful operations
Solution Approach 1:
Failed business object instances are extracted and removed from the batch job after identification. This allows the remaining successful instances to be saved without requiring resubmission of the entire batch, eliminating waste of time reprocessing successful operations
Solution Approach 2:
The system performs preliminary error detection and filtering of failed instances before saving successful ones. By identifying and removing failed instances early in the process, the system prevents unnecessary resubmission and reduces time loss
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
A service consumer may define batch jobs (batch containers) in which business object methods can be invoked on business object instances. The invocations may be recorded. The service consumer may trigger batch execution to cause the business object instances to be modified in accordance with the recorded invocations. The batch job can be executed as a single transaction in a single process. The batch job can be partitioned into multiple transactions and processed by respective multiple processes.