Optimize Performance
You can optimize performance in one of the following ways:
Load schemas into memory - Using this option, users can load all of their data into Incorta's memory. This can be done from any table's definition page, by disabling the Performance Optimized option.
Load schemas partially into memory - using the Load Filter.
Loading schemas in the staging area - This option This can be done by either Loading tables in the staging area or Selecting tables within a schema to load into memory.
Performance Testing
To determine benchmarks for various instances, Incorta ran two tests (a simple aggregated query against the number of records in the table), six times, on two different AWS machines using the following parameters:
- Turned off the cache so no data caching occured.
- Tested on one table with 2 or 3 billion records.
- Source data was in a .csv file
- Test did not change based on the AWS instance type
- Query: Group revenue by year and month
Test One
- Log into Incorta.
- Select Schema with one table and 2 billion records for the table.
- Run an aggregated query to show revenue by year and month.
Test Two
- Log into Incorta.
- Select Schema with one table and 3 billion records for the table.
- Run an aggregated query to show revenue by year and month.
The following table displays the results of the benchmark tests. The number of GB is the size of the data after it has been ingested into Incorta.
Machine types | 2 Billion Records (29.8 GB) | 3 Billion Records (41.91 GB) |
---|---|---|
AWS Win Machine m4-16x Large 256 GB RAM i-79f9fe61 | 2.46 to 2.58 seconds | 4.43 to 4.51 seconds |
AWS Linux Machine x1-32x-large | 1.28 to 1.39 seconds | 1.89 to 2.14 seconds |