An Improved Method for Performance Analysis of Generalized Integrated Interleaved Codes

Generalized integrated interleaved (GII) codes show great potential in communication and distributed storage systems for their good error-correction capability and locality. To design GII codes with optimal code parameters, an accurate theoretical performance analysis method is essential, which also benefits the study of decoding behavior. However, the current performance analysis method does not fit simulation results very well and can only provide frame-error-rate (FER) evaluation. This letter presents an improved theoretical performance analysis method, including both bit-error-rate (BER) and FER estimations, which fit simulation results very closely. More importantly, the proposed method could predict the failure probability of each decoding round, thus facilitating a better understanding of decoding behavior. To simplify the analysis, an approximation performance evaluation method is also introduced by relaxing the constrained conditions, which still exhibits similar accuracy.