Measurements-based modeling of burst errors on multiple parallel storage channels

This paper presents a measurement-based burst error model for storage devices that use multiple parallel channels when they are affected by external disturbances. The burst errors are modeled using a set of Markov processes and channel error measurements are exploited to specify the parameters of the Markov processes and to determine the correlation of errors in the various channels. The application of the proposed model to AFM-based probe storage devices is presented and numerical results for various cases of external noise sources are given.