Load Balanced Query Evaluation in Shared-Everything Environments

In this paper, we present data threaded execution, a new strategy to exploit both, pipelining and intra-operator parallelism in shared-everything environments. Data threaded execution is intuitive, straightforward to implement, but resistant against workload estimation errors and resistant against the discretization error of processor scheduling, that conventional strategies suffer from. Furthermore, data threaded execution minimizes startup and shutdown execution delays. Simulation results show that data threaded execution outperforms conventional strategies significantly due to the better utilization of parallel processing resources.