Bulk data provider using user defined or stock SQL statements. Some challanges
- Security
- Performance of JSON serialization
- Blocking execution
- No catalog of data (Governence)
- Async execution and result set saving prevents from multiple execution or cursor state management.
- gRPC based arrow based engine is way faster in large results
- Tags & Wiki allows end user programs to access metadata information
- *[Enterprise Only] You can expose dependency between
source -> dataset -> virtual dataset -> virtual dataset
- Authentication & Authorization is the same with Dremio UI.