Because of the query requirements, we are required to replicate a table 3 times with differing clustering keys positions. For example: table1 ( A PK, B CK, C CK, D CK), table2 (A PK, C CK, B CK, D CK), table3 (A PK, D CK, B CK, C CK). The table size is very large to replicate 3 times. We thought if we changed the data model to DSE Graph, we don't need to replicate the table. We could split this table into 2-3 small vertexes and set up edges to provide the necessary connections between them. Is our thought correct? Can DSE graph eliminate this replicate issue we have on Cassandra?