A File System for Information Management

Abstract

Nebula is a file system that explicitly supports informationmanagement. It differs from traditional systems in three important ways.First, Nebula implements files as sets of attributes. Each attributedescribes some property of the file such as owner, protection, functionsdefined, sections specified, project, or file type. The content of thefile is specified by a special text attribute. Second, Nebula supportsassociative access to a subset of files in one or more file systems.Finally, Nebula replaces traditional directories with database views. Adirectory implements a naming closure; i.e., the name of a file isassigned by the directories in which it resides. A view is a query thatselects file objects from a scoped index. Thus, the name of a file is aproperty of the file, not the view. Like directories, views can be namedand included in other views. When a file (or view) is created, it isplaced in an index, not a directory. The file appears in the appropriateset of views when the file system is refreshed. In this way, viewsdynamically classify files. The unique properties of the Nebula filesystem enable users to locate, classify, and identify informationquickly and easily as demonstrated by several Nebula-based applications.-- Nebula is unique in that it integrates the functionality of filesystems with the support provided by information services. The Nebulafile system explicitly supports information management across awide-area file system. Nebula is different from traditional systems inthree important ways. First, Nebula implements files as structured,typed file objects. Nebula implements these as a set of attributes. Eachattribute describes some property of the file such as owner, protection,functions defined, sections specified, project, or file type. Second,Nebula provides associative access to files within a scoped index.Finally, Nebula replaces traditional directories with views, whichdynamically classify information according to the needs of the user. Aview is a query that selects objects from a scoped index. Nebula viewsprovide a very powerful new file location technique that combinesbrowsing and searching; we use well-known and well understood browsingand searching commands like cd, ls, and cat.