Efficient file renames using B-tree based persistence for file system name spaces
Abstract:
Embodiments for optimizing file lookups after a file or directory rename operation. Files are defined using inode, dirent, and hash entries, and the name space of the files are stored in a B−Tree using file system conventions in which the inode, dirent, and hash values are referenced by the directory first followed by the inode number, as in a first data structure denoted [directory:inode]. Upon a rename of a file, the old inode, dirent, and hash entries are deleted, and a reference is created to the old name using the inode number first followed by the old directory, as in a second data structure denoted [inode:old_directory], and this reference is used as a forward link to the new file under the new directory name.
Information query
Patent Agency Ranking
0/0