Query throughput: Difference between revisions

Content deleted Content added

Inline

Latest revision as of 20:07, 13 February 2024

Redirect to:

@@ Line 1: / Line 1: @@
+#REDIRECT [[Transactions per second]]
-'''Query Throughput.'''  See [[Transaction Processing Performance Council]] (TPC).
+{{Rcat shell|
-The throughput metric is a classical throughput measure characterizing the ability of the system to support a multi-user workload in a balanced way. A number of query users (S) is chosen, each of which execute the full set of 17 queries in a different order. In the background there is an update stream that runs a series of insert/delete operations (one pair for each query user). The choice of the number of users is at the discretion of the test sponsor.
+{{R to related topic}}
+}}
-The throughput metric is computed as the total amount of work (S*17), converted to hours from seconds (3600 seconds per hour), scaled by the database volume (SF) and divided by the total elapsed time (Ts) required between the first query starting and the last query or update function completing.
-Therefore the complete formulation is:
-S * 17 * 3600
-QthD = --------------- * SF
-Ts
-Note that if the test sponsor chooses S=1 for the throughput test, for ease of benchmarking it is permissible to omit the throughput test and compute the throughput metric using timings obtained during the power test. Similarly, it is permissible to schedule the insert/delete activity for the throughput test after all the queries complete.