The service class controls query prioritization. Service classes define how an Aster system divides the available hardware resources among all active queries and other activities, such as physical backups, in the system.
When you create a service class for an Aster system, you can set the processing priority and memory limits.
- Processing priority represents the overall importance of the workload compared to other workloads. Weight is the relative resource allocation of the CPU and disk I/O within the priority. The weight value dictates the ratio of resource allocation. For example, if two statements execute with the same priority, but with weight values of 80 and 20, the system will aim to allocate resources in a 4:1 ratio, with most of the resources allocated to the statement with higher weight.
- Memory limits allows you to specify the maximum total memory per node for all queries associated with this service class. Memory soft limit is the highest percentage of memory the workload can occupy when there is contention for memory in the system. Memory hard limit is the highest percentage of memory the workload can occupy even if there is no contention for memory in the system.
These settings apply to all queries that map to that service class. Each service class can be used by one or more workloads.
Reading more about these limits in the Teradata Aster® Database User Guide is recommended before creating service classes in Teradata Viewpoint.