Understanding Data Movement in Parallel Data Warehouse

To fully leverage the power of the Parallel Data Warehouse (PDW), data is distributed across all of the nodes within the appliance. Ideally, the data that each node requires will be located on that node. Unfortunately, there are a number of situations where this isn't the case. To solve this, PDW sometimes moves data between nodes within the appliance. At times this movement is desired and at other times, it is not, resulting in possible performance issues. In this session, we'll review the different types of data movement within the PDW, explaining the need and benefit of each type of movement. At the end of the session, you'll understand each type of data movement and understand how to identify when the data movement isn't desirable with options on how it can be mitigated.