Tripling Performance by Using Thread per Core with DataStax Enterprise
In this talk we’ll explore migrating a 10TB Oracle Rack to a DataStax Enterprise cluster. During this process, many data models were considered including tables per year and whether to store table data as a wide column or BLOB (serialized by avro pojo). After upgrading to DSE 6.0 we updated our data models to leverage the new Thread Per Core architecture resulting in a triple performance increase from previous models by querying different shards.