Disk Failure Prediction Model for Information Systems based on SMART Technology

This paper proposes disk failure prediction model using disk SMART technology for information systems, which is based on the actual data, threshold and attribute value of the SMART parameters to establish parameter reliability model, through theoretical discussion and practical test and analysis to choose the weight of important parameters, according to the parameter reliability model and the parameter weights to create disk reliability model. Final, disk reliability failure threshold is determined by the analysis the SMART parameter actual data and parameter weights case, combined with the disk reliability model and disk reliability failure threshold to classify the level of disk reliability. Experimental results show that this technology can satisfies the actual usage and achieves the prediction result.