Post-layout estimation of side-channel power supply signatures

Two major security challenges for integrated circuits (IC) that involve encryption cores are side-channel based attacks and malicious hardware insertions (trojans). Side-channel attacks predominantly use power supply measurements to exploit the correlation of power consumption with the underlying logic operations on an IC. Practical attacks have been demonstrated using power supply traces and either plaintext or cipher-text collected during encryption operations. Also, several techniques that detect trojans rely on detecting anomalies in the power supply in combination with other circuit parameters. Counter-measures against these side-channel attacks as well as detection schemes for hardware trojans are required and rely on accurate pre-fabrication power consumption predictions. However, available state-of-the-art techniques would require prohibitive full-chip SPICE simulations. In this work, we present an optimized technique to accurately estimate the power supply signatures that require significantly less computational resources, thus enabling integration of Design-for-Security (DfS) based paradigms. To demonstrate the effectiveness of our technique, we present data for a DES crypto-system that proves that our framework can identify vulnerabilities to Differential Power Analysis (DPA) attacks. Our framework can be generically applied to other crypto-systems and can handle larger IC designs without loss of accuracy.