Requests in an SLG tier workload divide resource allocations equally. For example, if the workload is allocated a 5% share and there are five requests running concurrently, each request gets 1% of the shared resources.
To accommodate growth, Teradata recommends that you assign workloads a higher allocation percentage than you think is needed. If workload consumption varies over time, allocate the workload a different percentage for each planned environment. Do not relocate workloads to different SLG tiers in different planned environments.
If there is more than one SLG tier, the allocation percentages assigned to workloads on the higher tiers translate to a larger share of resources than those on the lower tiers. For example, SLG tier 1 workloads receive service ahead of SLG tier 2 workloads.
The unused resources flowing down from an SLG tier to the lower tiers equals 100% minus the sum of the workload allocation values on that tier. In the preceding figure, the sum of the three workload allocation values on the tier is 30%. This guarantees that at least 70% of the resources flowing into that tier are available to workloads in the lower tiers.
See the Orange Book, Teradata Priority Scheduler for Linux SLES 11, 541-0008867 for guidance on how to set up SLG tier workloads.