HammerDB evaluation format

[perdelt]

Input: Text file logs

• configuration, e.g. MariaDB-BHT-10-9-4
• experiment run, e.g. 1
• benchmarking instance, e.g.
  • number in sequence
  • number of pod (process)
  • number of vuser (thread) inside pod

Result: Table (pandas DataFrame)

• Index: Unique name of benchmark
• Columns: configuration, experiment run, number in sequence, number of pod, number of vuser, metrics (NOPM and TPM)

Also result: Table (pandas DataFrame) per pod

Tool: Aggregate

• here: only summation (?)
• per number in sequence of a configuration experiment run

here: no loading information

[perdelt]

Open: Aggregation, test if vusers are correct