The following performance improvements have been made since Release 9.0 of Connect.
| Release | Applications with Improved Performance |
| 9.10.33 | Large number of source files in a copy task. The elapsed time improves significantly for a Connect-ETL task that processes a large number of source files with the “Original order of source records need not be maintained” performance tuning option set |
| 9.9.20 | Task Editor Target Database Table dialog. When developing tasks in the Task Editor, mapping, unmapping, and editing target column mappings in the Target Database Table dialog are now more responsive. |
| 9.9.6 | Hive and Kafka Sources and Targets. The memory usage improved significantly for Connect for Big Data jobs that has multiple HIVE or Kafka sources and/or targets. |
| 9.7.34 |
Apply CDC Changes to Impala Targets. The elapsed time for CDC apply changes tasks is significantly improved when applying changes to Apache Impala tables using the Kudu storage format. |
| 9.7.23 | Kudu Targets. The elapsed time improves for Connect for Big Data Intelligent Execution jobs having kudu tables connected through JDBC connections. |
| 9.7.23 |
Hive and Impala Targets. The elapsed time improves for Connect for Big Data Intelligent Execution jobs running on the cluster having Hive/Impala tables connected through JDBC connections. |
| 9.7.19 |
Connect Change Data Capture for DB2 z/OS. Improved replication performance on low-volume DB2 tables. Refer to the “Flushing DB2 Log Buffer to Reduce Data Capture Delays” section in the DB2 z/OS Change Data Capture Reference |
| 9.7.1 |
Hive and Impala Sources. The elapsed time improves for Connect for Big Data Intelligent Execution jobs running on the cluster having Hive/Impala tables connected through JDBC connections |
| 9.4.6 |
Hive sources. The elapsed time improves for jobs having Hive sources connected through a JDBC connection. |
| 9.3.16 |
Hive sources and targets. The elapsed time improves for Connect for Big Data Intelligent Execution jobs running on the cluster having multiple Hive sources or targets connected through JDBC connections defined in the same task or in different tasks. |
| 9.1.6 |
Connect for Big Data MapReduce jobs. Connect for Big Data’s updated algorithm for calculating the number of reducers in MapReduce jobs should improve performance. |
| 9.1.2 |
Connect for Big Data Spark jobs. Connect for Big Data’s updated algorithm for calculating the number of reducers in Spark jobs should improve performance. Additionally, in a multi-tenant scenario, if Spark's dynamic allocation mechanism is enabled, Connect for Big Data will use resources more appropriately. |
| 9.0.7 |
Connect for Big Data Spark jobs. Both elapsed time and load balancing when Spark outputs to local disk have been improved. |
| 9.0.3 |
Hive targets. When writing a large number of records to Hive, the elapsed time may improve if constant values are mapped to partition columns. |