Analyzing power and energy consumption of large join queries in database systems

Performance of large join queries has been widely addressed in the database literature, explicitly the problem of finding a join order that minimizes the execution cost of a given query has been treated. Lately, the elevated electricity consumption of data center facilities have lead to the development of energy-efficient hardware and software. Relational Database Management Systems (RDBMS) have been redesigned to predict power consumption of queries and guide plan selection towards energy reduction goals. In this article, a power consumption comparison between different large join query optimization approaches will be presented. For that purpose, three large join query optimization algorithms will be used, the PostgreSQL genetic algorithms GEQO, a simulated annealing approach SAIO and an automata theory based meta-heuristic DSQO our previous work. The main goal of our research is to analyze the power behavior of different large join query optimizers solving queries derived from the TPC-DS benchmark.