Identification of FSM State Registers by Analytics of Scan-Dump Data

Big data analytics have gained tremendous successes in mining valuable information various fields. However, its potential to solve complex problems hardware security has not been adequately tapped. This paper presents a non-invasive approach identify the state registers of finite machine (FSM) an integrated chip. The FSM are mined from scan-dump by exploiting strongly connected property and chronologically correlated codes FSM. sequence scanned out each scan register is partitioned into non-overlapping strings high weighted frequencies string-matching algorithm. A coherency between pair defined computed based on strings. dimension matrix first reduced pruning some low influence regression analysis. then clustered minimize within-cluster variances their values. proposed scheme applied IP cores OpenCores. experimental results show that our can correctly most designs with hit rate.