Graph storage in Cassandra keyspace and tables

Describes graph storage in Cassandra at a high level.

DSE Graph uses Cassandra to store schema and data. Three Cassandra keyspaces are created for each graph, <graphname>, <graphname_system>, and <graphname_pvt>. For example, for a graph called food, the three keyspaces created will be food, food_system, and food_pvt. The first keyspace food will hold the data for the graph. The second keyspace food_system holds schema and other system data about the graph. The third keyspace food_pvt holds information about partitioning for vertices should the graph contain a vertex with a large number of edges that requires the vertex table to be partitioned across the cluster.

In the <graphname> keyspace, two tables are created for each vertex label to store vertex and edge information, vertexLabel_p and vertexLabel_e, respectively. For example, for a vertex label author, two tables are created, author_p and author_e.