As memory scales down to smaller technology nodes, new failure mechanismsemerge that threaten its correct operation. If such failure mechanisms are notanticipated and corrected, they can not only degrade system reliability andavailability but also, perhaps even more importantly, open up securityvulnerabilities: a malicious attacker can exploit the exposed failure mechanismto take over the entire system. As such, new failure mechanisms in memory canbecome practical and significant threats to system security. In this work, we discuss the RowHammer problem in DRAM, which is a prime (andperhaps the first) example of how a circuit-level failure mechanism in DRAM cancause a practical and widespread system security vulnerability. RowHammer, asit is popularly referred to, is the phenomenon that repeatedly accessing a rowin a modern DRAM chip causes bit flips in physically-adjacent rows atconsistently predictable bit locations. It is caused by a hardware failuremechanism called DRAM disturbance errors, which is a manifestation ofcircuit-level cell-to-cell interference in a scaled memory technology. Weanalyze the root causes of the RowHammer problem and examine various solutions.We also discuss what other vulnerabilities may be lurking in DRAM and othertypes of memories, e.g., NAND flash memory or Phase Change Memory, that canpotentially threaten the foundations of secure systems, as the memorytechnologies scale to higher densities. We conclude by describing andadvocating a principled approach to memory reliability and security researchthat can enable us to better anticipate and prevent such vulnerabilities.
展开▼
