Details
-
Improvement
-
Status: Closed (View Workflow)
-
Minor
-
Resolution: Done
Description
As a user of Pipeline, it's frustrating that it writes so much redundant data for each FlowNode. This is both disk-space inefficient and increases the amount of data read/written to disk (more disk throughput). We should use XStream aliases for the most common and verbose objects at the very least to reduce this some.
Implemented with workflow-support 2.17 – we used aliases for the Tag objects of the storage engines, the packages of common Actions, and the package with FlowNode implementations.
The result is to reduce size-on-disk (and the amount of data written) by about 1/3.