Apache Cassandra - Clustering

Clustering

When the cluster for Apache Cassandra is designed, an important point is to select the right partitioner. Two partitioners exist:

1. RandomPartitioner (RP): This partitioner randomly distributes the key-value pairs over the network, resulting in a good load balancing. Compared to OPP, more nodes have to be accessed to get a number of keys.
2. OrderPreservingPartitioner (OPP): This partitioner distributes the key-value pairs in a natural way so that similar keys are not far away. The advantage is that fewer nodes have to be accessed. The drawback is the uneven distribution of the key-value pairs.