Evaluation of optimization control parameters in multi-leveled cloud platform

摘要

Cloud platforms demand the highest performance available. To achieve this different optimization methods were proposed and tested. Each method has its own set of parameters and settings. While consolidating them into one single hierarchy leads to increase in performance, this consolidated method is also more complex. Thus, it is almost impossible to foresee which parameters should be configured and which values they should have without a real testing environment. However even with such an environment a set of experiments need to be performed. Those experiments provide comparison data, allowing determine values for such parameters. In this paper such experiments were performed and values were gathered which will help to properly configure the consolidated method and gain increase in performance for the system.