The Scientific World Journal

Research Article

Evolutionary Multiobjective Query Workload Optimization of Cloud Data Warehouses

Multiobjective optimization of Cloud database configuration using branch-and-bound.

Input: Set of VM types (VMs), Query workload ()
Output: A set of pareto-optimal solutions
() QP: Query Plan
() Q: Queue of QPs
() Calculated_value: Multiobjective cost of optimization
() S: Set of solutions
() for ( to all configurations of (VMs)) do
() B_response_time Find_the_best_response_time (, )
() Cheapest_cost Find_the_cheapest_cost (, )
() Heuristic_value Calculate_heuristic_value (B_response_time, Cheapest_cost)
() Q = null;
() ;
() for (Each Query Plan QP in the workload ) do
() Q.Enqueue QP_in();
() Calculated_value = Calculate_with (, Q);
() if (Calculated_value is worse than Heuristic_value}) then
() Break the loop and start with the next VM configuration;
() if (is empty and the Calculated_value is better than heuristic_value) then
() Add () to the solution set of ;
() return ;