Evaluation Plan for SIGMOD2021

[Y-Wakuta]

Workload

• TPC-H
• RUBiS

methods to compare with

• Static optimization that only has materialized view plan for all queries
• NoSE result for the start time of time_depend workload
• NoSE result for the end time of time_depend workload
• time_depend schemas created by professional (maybe ask zenmyo-san to create this is good)
• Proposed method

setting choices

Although the time_depend optimization part is the key part of this research, pruning techniques are also an important part. One possible connection of these two contributions is 'pruning is required because the model becomes more complicated to apply to time_depend workload'.

• evaluate optimization time length and schema accuracy with each pruning method compared to the non-pruning pattern.
• If the optimization can not be executed without pruning, describe that clearly by showing like 'the optimization without any pruning method could not be done with 24 hours'

Evaluating scalability changing the number of servers is also important.

[Y-Wakuta]

I could not make it in time for SIGMOD