Database as Filesystem

C1 | Tue 22 Jan | 11:05 a.m.–11:30 a.m.

Presented by

  • Peter Chubb
    http://ts.data61.csiro.au/people/?cn=Peter+Chubb

    Peter has been a long-term contributer to open source (his first patch was to international iSpell in around 1990 to enable Australian spelling rules), to Unix (he worked on the Unix kernel for SGI, Fujitsu, and for AT&T Bell Labs while at Softway Pty Ltd), and to Linux systems software (kernel and low-level software like u-boot and qemu). He has spoken at many LCA events. Peter has never used a Windows operating system except when forced to, and then for only a short time.

Abstract

Most network file systems are either a layer over an existing filesystem (NFS, CIFS), or are develped from scratch to have separate, replicated, purpose-designed databases for metadata and object store (GFS, Glusterfs). At the same time, most database engines provide (or can be coerced into providing) replication and all the ACID properties needed for a high-performance filesystem. Idea: Use a database engine (Postgres, MariaDB) on raw partitions with a fast separate nVME log file; build POSIX file system semantics on top. It's pretty obvious that this could work; I'm just starting to implement it so performance and durability can be measured.