ISSN 2063-5346
For urgent queries please contact : +918130348310

Hybridization of KF and GA as a Predictive Algorithm for Reduction of Power Consumption in a Cloud Computing Data Centre

Main Article Content

Afolabi Rotimi, Bamidele Adebisi, Anthony U. Adoghe
» doi: 10.31838/ecb/2023.12.si6.262


Data Centre (DC) has becoming a major and important component of a cloud computing to meet up with the rapid increase in the demand of telecommunication services. However, the cost of maintaining a DC is very high due to high power consumption of the unit in a telecommunication industry. The situation is further exacerbated in a country like Nigeria where there is highly unstable power supply from the national grid. Kalman Filter (KF) which is one of the energy consumption optimization technique through power consumption prediction model used to reduce the power consumption is characterized with high prediction error due to random selection of KF parameters’ value during the prediction period. Hence, in this paper, Hybridization of KF and GA for power reduction through accurate power consumption prediction is proposed. Data were collected from Four different servers in Nigeria, named BSC 13, BSC 14, RNC 05 and RNC 06 using power analyzer, multimeters and thermometer. The historical assessment of data collected were carried out for the DC for two years (January to December of 2019 and 2020). The GA was used to obtain best possible values for the KF parameters and KF was then used to predict the future power consumption value on hourly basis for each day of the week. The proposed PCoKFGA model gave better performance with accurate prediction, higher power usage effectiveness and lower energy consumption than the existing KF model. Therefore, the PCoKFGA model proposed would be most useful where accuracy in prediction is of utmost importance, and for run-time application.

Article Details