Details emerge on next-gen Windows file system
Posted 10/17/2003 @ 11:54 AM, by Ken "Caesar" Fisher
File systems may not be the center of many water cooler conversations, but Microsoft's next-generation plans have been the subject of much rumor mongering and guesswork amongst Windows enthusiasts. Currently, the NT File System (NTFS) rules the roost, providing the storage foundation to NT, Windows 2000 and XP (with slight changes, here and there). However, there's talk of a next-generation technology called WinFS, which some people expected would totally replace the now-aging NTFS. It has become clear that WinFS is instead going to be a kind of database qua filesystem solution that will be built atop NTFS.
NTFS is only one component of the revamped storage system in WinFS. Another key building block is the querying capabilities of Microsoft's SQL Server relational database, according to Microsoft. WinFS also will incorporate the data labeling capabilities of Extensible Markup Language (XML), Muglia said. "Think of WinFS as pulling together relational database technology, XML database technology, and file streaming that a file system has," he said. "It's a (storage) format that is agnostic, that is independent of the application."
Microsoft is taking a step in the right direction here. File systems are perhaps the most antiquated feature of modern OSes. Although there have been rare innovations such as BeOS' journaling file system, most solutions seem rather archaic when compared to the other innovations in computing in the past decade. Moving to a more database-like solution should make data storage, retrieval, searching, and linking much faster and more convenient. Still, much of the convenience will rely on the use of metadata-that is, data about data-to describe files, but as of yet there's been no sure fire way to get users to enter or manage metadata. You can already see this by analogy with most people's MP3 collections. Most users rely on filenames, e.g., 07-Rockin the Suburbs.mp3, to identify a song or artist. The more ambitious of us create hierarchies of directories to provide more information, and to make things easier to find (e.g., C:\Music\Rock\B\Ben Folds\Rockin the Suburbs\07-Rockin the Suburbs.mp3). But if you've ever used a program like MusicMatch or iTunes, you also know that you can use MP3 ID tags to organize your library within those applications, without touching the actual file structure. The latter solution is the best one if a) you have all of your ID info filled in, and b) you only use one client to play your files (because the Library is built and maintained by only the app in question). What Microsoft wants to do is take metadata like MP3 IDs and make them accessible to all applications and services through the file system, for all kinds of data. Such a solution, if implemented the right way, would also let applications add their own metadata to files, without breaking them. So, for example, an MP3 player could store per-song EQ settings, play counts, and other non-ID tag information in the file without making the file unplayable.
That said, "built atop NTFS" leads to me to think that this is going to be more of an update to the indexing service than a change to the file system per se. However, more details should be available by the end of the month, so let's play wait-and-see.
p.s. Ben Folds rocks.
