Connection pooling

The DataStax drivers maintain a pool of connections to each of the nodes selected by the load balancing policy.

By default, the driver instance creates one connection to each of the local datacenter hosts in the default load balancing policies.

Connection pooling is separate from the initial contact points. Initial contact points are supplied to the driver instance. Those contact points are used only to establish the control connection to discover the Cassandra or DSE cluster topology.

Connection pools are accessed asynchronously. Multiple requests can be submitted on a single connection simultaneously. For most workloads, it is recommended to use one long-lived connection from the driver to each DSE server. If the default connection pool settings are not adequate, the number of connections per host and the maximum number of simultaneous requests per connection are configurable. The binary protocol allows up to 32,768 concurrent requests per connection.

For information about configuring connection pools, see the following:

The DataStax Python driver doesn’t support customized connections per host due to single thread/GIL.