Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.
DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.
Userspace library file systems (LibFSes) promise to unleash the performance potential of non-volatile memory (NVM) by directly accessing it and enabling unprivileged applications to customize their LibFSes. Unfortunately, such benefits pose a significant challenge to ensuring metadata integrity. Existing works either underutilize NVM’s performance or forgo security guarantees. We present Trio , a userspace file system architecture that resolves this inherent tension. Our key insight is that other state ( i . e ., auxiliary state) in a file system can be regenerated from its “ground truth” state ( i . e ., core state). Trio explicitly defines the data structure of a single core state and shares it as common knowledge among its LibFSes and the trusted entity. A LibFS directly accesses NVM without involving the trusted entity and can be customized with its private auxiliary state. The trusted entity enforces metadata integrity by verifying the core state of a file when its write access is transferred from one LibFS to another. We design a generic POSIX-like file system called ArckFS and two customized file systems based on Trio . Our evaluation shows that ArckFS outperforms existing NVM file systems by 3.1 × to 17 × on LevelDB while the customized file systems further outperform ArckFS by 1.3 ×.