One of the key features that enables the Greenplum Database to scale linearly and achieve such high performance is its ability to utilize the full local disk I/O bandwidth of each system. Typical configurations can achieve from 1 to more than gigabytes per second sustained I/O bandwidth per machine. This rate is entirely scalable, and overall I/O bandwidth can be increased linearly by simply adding nodes without any concerns of saturating a SAN/shared-storage backplane. In addition, no special tuning is required to ensure that data is distributed across nodes for efficient parallel access. When creating a table, a user can.
