A completely automated upgrading method for capacity of IT system resources is disclosed. Due to the adaptability of our methodology it does not require any manual intervention at any stage of prediction. The method comprises the steps of monitoring over time a signal representing the capacity of a predetermined IT system resource and collecting a dataset of said signal, then analysing said dataset through a prediction method to forecast said dataset behaviour and further upgrading said capacity at time t-n, through allocating additional resource to the IT system, when said estimated dataset behaviour shows that at time t said resource reaches a critical threshold, said prediction method being based on our automation of the Box-Jenkins method and Further comprising subsequently (a) to pre-process dataset, where missing values of the dataset are filled, level discontinuities and outliers are filtered out, the last P samples are left out, (b) to find a trend of the remaining dataset, where trend is identified and fiitered out of the dataset through coefficient of determination methodology, (c) to detect seasonality through computing an auto correlation function of the remaining dataset, further repeating the same detection on aggregate series of the previous dataset, and removing detected seasonality through a seasonal differencing process, (d) to model the resulting stationary series under an ARMA process, and said prediction method to forecast said behavior at time t being computed following the inverse order of steps (b)-(d).

Automated upgrading method for capacity of IT system resources

CREMONESI, PAOLO;DHYANI, KANIKA
2010

Abstract

A completely automated upgrading method for capacity of IT system resources is disclosed. Due to the adaptability of our methodology it does not require any manual intervention at any stage of prediction. The method comprises the steps of monitoring over time a signal representing the capacity of a predetermined IT system resource and collecting a dataset of said signal, then analysing said dataset through a prediction method to forecast said dataset behaviour and further upgrading said capacity at time t-n, through allocating additional resource to the IT system, when said estimated dataset behaviour shows that at time t said resource reaches a critical threshold, said prediction method being based on our automation of the Box-Jenkins method and Further comprising subsequently (a) to pre-process dataset, where missing values of the dataset are filled, level discontinuities and outliers are filtered out, the last P samples are left out, (b) to find a trend of the remaining dataset, where trend is identified and fiitered out of the dataset through coefficient of determination methodology, (c) to detect seasonality through computing an auto correlation function of the remaining dataset, further repeating the same detection on aggregate series of the previous dataset, and removing detected seasonality through a seasonal differencing process, (d) to model the resulting stationary series under an ARMA process, and said prediction method to forecast said behavior at time t being computed following the inverse order of steps (b)-(d).
File in questo prodotto:
File Dimensione Formato  
Initial Publication.pdf

Accesso riservato

: Altro materiale allegato
Dimensione 2.26 MB
Formato Adobe PDF
2.26 MB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11311/589894
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact