New storage-class memory (SCM) technologies, such as phase-change memory, STT-RAM, and memristors, promise user-level access to non-volatile storage through regular memory instructions. These memory devices enable fast user-mode access to persistence, allowing regular in-memory data structures to survive system crashes.
Mnemosyne provides a simple interface for programming with persistent memory. Mnemosyne addresses two challenges: how to create and manage such memory, and how to ensure consistency in the presence of failures. Without additional mechanisms, a system failure may leave data structures in SCM in an invalid state, crashing the program the next time it starts.
In Mnemosyne, programmers declare global persistent data with the
persistent or allocate it dynamically. Mnemosyne provides
primitives for directly modifying persistent variables and supports
consistent updates through a lightweight transaction
mechanism. Compared to past work on disk-based persistent memory,
Mnemosyne is much lighter weight, as it can store data items as small
as a word rather than a virtual memory page. In tests emulating
the performance characteristics of forthcoming SCMs, we find that
Mnemosyne provides a 20--280 percent performance increase for small
data sizes over alternative persistence strategies, such as Berkeley
DB or Boost serialization that are designed for disks.