Types of Batch
1. Normal batch
This is a batch program that is periodically or non-periodically executed at a specified time by a Job Scheduler, and is basically executed by the Shell (runJob.sh or common Shell) provided by the framework.
The types of batch jobs developed as normal batch are as follows (may vary depending on the project performed/business characteristics):
-
Self Check
-
Initialization tasks
-
Closing tasks
-
Report tasks
-
Mass data modification
-
Data creation/aggregation and writing to a separate Table, Sam file generation
-
Data Convert
-
Tasks such as error handling/correction/cancellation
-
File/message creation followed by transmission and data reception, etc.
-
Programs for ledger Save/Load, Recovery, Dump, etc.
2. Ondemand batch
This is a batch program that executes a batch using an API from an online service when large-volume data processing is required not by a Job Scheduler. The batch is executed asynchronously (Async) from the online service.
The types of batch jobs developed as ondemand batch are as follows (may vary depending on the project performed/business characteristics):
-
Batches that complete within several minutes (Short Term Job) (targeting batches whose processing time is longer than online processing time but shorter than the time required for normal batch processing)
-
Cases where a batch must be executed directly from an online program
3. Daemon batch
This is a batch program that runs automatically at regular intervals according to the schedule time while it is started in the form of a daemon program.
The characteristics of daemon batch are as follows:
-
Development of daemon batch jobs is the same as writing normal batch job modules
-
Control is possible according to system conditions while the batch is executing
-
Multi-node processing function
-
Schedule time definition
The types of batch jobs developed as daemon batch are as follows (may vary depending on the project performed/business characteristics):
-
Resident daemon for monitoring
-
Resident daemon for business data correction/cancellation, etc.
-
Data reception, etc.
Since daemon batch operates as an individual process for each registered batch job, caution is required when registering jobs for daemon batch depending on server resource conditions (CPU/memory, etc.).