A forest, many species of tree
Reading linux filesystem challenge and noting (or not) the responses from the slashdot community. The author starts with the emphasis on data integrity, recovery, and fault tolerance.
He mentions some of the various journalling filesystems under linux; ReiserFS, ext3, JFS and XFS.
Here the article changes and looks at Apple's and Microsoft's upcoming filesystems (Tiger(?) and longhorn. Both of these are more database like, organising by attribute, metadata describing the type of file etc. The author looks for something similar for linux in it's future. Nowhere further is data integrity, recovery, or fault tolerance mentioned. Does a 'more database like filesystem' automatically include these features ? Or can we have the integrity, recovery, and fault tolerance to that of a filesystem like Solaris's upcoming ZFS ? This together with nautilus (possibly improved) to handle the organisation etc of one's files could see a more scalable and robust grouping of requirements such as fast searching and data integrity of one's data.
Posted at 11:58PM Jul 28, 2004 by smg in General |