RowScanner: Identification and Prediction of Rowhammer Attacks on DDR3/4

By: Raj Gautam Dutta and Yier Jin

Stage: RTL

Summary

• An innovative framework that can identify, predict, and help to mitigate RowHammer attack based on an analytical model of capacitive-coupling vulnerabilities in DDR3 and DDR4 DRAMs.
• The framework can achieve two goals: (i) predict whether a given memory chip is vulnerable to a RowHammer attack, thereby enabling the installation of resilient DRAMs in computing systems and (ii) accurately locate vulnerable cells within the DRAM chip if present.
• The framework was used to validate the presence/absence of vulnerable memory cells in DRAMs of different manufacturers.

Contact

Yier Jin

Acknowledgments

• This work was supported in part by Department of Energy (DOE).