Two Stages Query Processing Optimization Based on ELM in the Cloud

Abstract

As one variant of MapReduce framework, ComMapReduce adds the lightweight communication mechanisms to improve the performance of query processing programs. Although the existing research work has already solved the problem of how to identify the communication strategy of ComMapReduce, there are still some drawbacks, such as relative simple model and too much user participation. Therefore, in this paper, we propose a two stages query processing optimization model based on ELM, named ELM to ELM (E2E) model. Then, we develop efficient sample training strategy, predicting and execution algorithm to construct the E2E model. Finally, extensive experiments are conducted to verify the effectiveness and efficiency of the E2E model.