Title :
Hourly server workload forecasting up to 168 hours ahead using Seasonal ARIMA model
Author :
Tran, Van Giang ; Debusschere, Vincent ; Bacha, Seddik
Author_Institution :
G2Elab - Grenoble Electr. Eng. Lab., INP - France, France
Abstract :
Data center workload prediction is important to take decisions in resources management system. Seasonal ARIMA model provide a good server workload methodology for the server workload forecasting. A large set of our experiments confirm that it has high performance, scalability and reliability and will bee integrated in our system. This paper presents a general expression in development of our forecast model in the project EnergeTic-FUI, France.
Keywords :
autoregressive moving average processes; computer centres; energy management systems; file servers; EnergeTic-FUI; data center workload prediction; forecast model; hourly server workload forecasting; resources management system; seasonal ARIMA model; server workload methodology; Atmospheric modeling; Biological system modeling; Knowledge engineering; Predictive models; Reliability engineering; Support vector machines; EnergeTIC-FUI; Seasonal ARIMA; data center workload forecasting; server workload;
Conference_Titel :
Industrial Technology (ICIT), 2012 IEEE International Conference on
Conference_Location :
Athens
Print_ISBN :
978-1-4673-0340-8
DOI :
10.1109/ICIT.2012.6210091
