DSE In-Memory is a part of the multiple storage options offered in DataStax Enterprise for optimizing performance and cost goals. DSE in-memory provides lightning-fast performance for read-intensive situations. It allows developers, architects, and administrators to easily choose what parts (some or all) of a database reside fully in RAM. It is designed for use cases that lend themselves to in-memory computing, while allowing disk-based workloads to be serviced by DSE Tiered Storage and traditional storage modeling.
DSE In-Memory is suitable for use cases that include primarily read-only workloads with slowly changing data and/or semi-static datasets, such as a product catalog that is refreshed nightly, but read constantly during the day. It is not suitable for workloads with heavily changing data or monotonically growing datasets that might exceed the RAM capacity on the nodes/cluster.
DataStax recommends using OpsCenter to check performance metrics before and after configuring DSE In-Memory.
- Creating or altering tables to use DSE In-Memory
Use CQL directives to create and alter tables to use DSE In-Memory.
- Verifying table properties
cqlsh, use the
DESCRIBEcommand to view table properties.
- Managing memory
How to manage available memory when using DSE In-Memory.
- Backing up and restoring data
The procedures for backing up and restoring data is the same procedure for DSE In-Memory data and on-disk data.