Towards a Highly Adaptable Filesystem Framework for Linux
Linux is growing richer in independent general purpose file systems with their own unique advantages, however, fragmentation and divergence can be confusing for users. Individual file systems are also adding an expanding number of options (e.g. ext3) and variations (e.g. reiser4 plugins) to satisfy new requirements. Both of these trends indicate a need for improved flexibility in file system design to benefit from the best of all worlds. We explore ways to address this need, using as our basis, KFS (K42 file system), a research file system designed for fine-grained flexibility.
KFS aims to support a wide variety of file structures and policies, allowing the representation of a file or directory to change on the fly to adapt to characteristics not well known a priori, e.g. list-based to tree-based, or small to large directory layouts. It is not intended as yet another file system for Linux, but as a platform to study trade-offs associated with adaptability and evaluate new algorithms before incorporation on an established file system. We hope that ideas and lessons learnt from the experience with KFS will be beneficial for Linux file systems to evolve to be more adaptable and for the VFS to enable better building-block-based code sharing across file systems.