In-Memory Data Collections (IMDC) are data stores that are available to users of the Aster SQL-MapReduce framework. IMDCs are:
- Structured
An IMDC has a schema.
- Distributed
A partition is associated with each parallel instance in the Aster Database MPP.
- In process
An IMDC resides in the JVM process that constitutes the Aster SQL-MapReduce framework. Therefore, the data is close to the Aster Analytics engine that uses it.
- In memory
An IMDC is in main memory, not on a constituent node of the Aster Database MPP.
- Snapshots of the data in a table on disk
If the data in the table changes, the data in the IMDC does not.
Using the IMDC API and SQL-MapReduce utilities, you can perform these operations:
- Create an IMDC
- Remove an IMDC
- Add data to an IMDC
- Iterate over data stored in an IMDC
- Get IMDC-related metadata*
* You can perform this operation during contract negotiation.