Lately there’s a lot of buzz about Windows File System (WinFS)
that’s raising several questions in our minds. What is it? When will
it be released? Where would it fit? and why should we even care?
This article will answer several questions about the upcoming WinFS
(at least what we know so far), explain its importance, and show you
how it can change the way we manage information on our computers
today.
Before we even begin our discussion, let’s make a couple of things
very clear. First, even though a lot of documents all over the
Internet state that WinFS will be a feature included in
Longhorn (code name for the next version of Windows client), that is
not going to be the case. Microsoft has said that WinFS will be
available on Windows platform at a future date. Everyone expected
WinFS to be a part of Longhorn for years but recently Microsoft has
made it clear that we should not count on WinFS to be released along
with Longhorn.
Secondly, the information provided in this article may change
because the storage system was not released to manufacturing at the
time of writing. Although the information in this article is based
on content that Microsoft has made publicly available, Microsoft
reserves the right to change the information at any time without
prior notice.
What is WinFS?
WinFS is the code name for a new type of file storage system that
will be part of future Windows platform after Longhorn is released.
The file system is revolutionary in the sense that it will allow us
to manage data in a way that has never been possible in the past. It
will change the way we think on how data should be stored. It will
also make access to information seamless across multiple
applications and formats.
Advantages of WinFS
One of the major problems that organizations face today is that they
have all the information stored across their organization in
disparate formats. Wouldn’t it be nice if we could search for
content on our corporate network independent of the format in which
the data was stored? That’s exactly the goal with WinFS. Users will
be able to search for content based on the metadata of the stored
item, so it wouldn’t matter whether the content was stored in Word,
Excel, PowerPoint, or a PDF file. The type of document and the
application that created the content will be irrelevant.
The list of types of data we store today on our computer is rather
large. We not only store standard text documents but also e-mail
contacts, music files, PDF documents, spreadsheets, graphics,
videos, and so on. The newer WinFS technology, which is implemented
using a relational database, will make it easier for us not only to
locate the information but also to transfer that information to
other stores on different computers. In addition, backups and
restores would be more efficient, not only because WinFS uses
database technology but also because of this ability to replicate
data to stores on other computers.
Overview of WinFS
Each one of us seems to be using a different method to keep
track of our data and over the years we have developed various
methods that facilitate locating our content in a more efficient
manner. For example, we find our own method of creating folders
and organize data in a way that gives us quick access to the
content. Let’s say you have hundreds of photographs from
vacations, birthdays, and other special occasions and you want
to organize them in folders in a way that they are separated yet
easy to find. One way to do that is to store them in a folder
that is labeled by the date, such that the folders are sorted
numerically. Here’s an example.
2004_08_14
2004_09_07
2005_01_15
2005_02_12
2005_03_04
The above example lists the folders in Windows Explorer so they
are sorted in a certain order, first by year, then by month, and
finally by the date. Similarly, we can label the files so they
are easy to locate. This is just one example. The items are
categorized in a certain way. Typically, your camera
software will name your files in a similar manner to make
your life a bit easier. However, even this cataloging has
limitations. Now you have tons of files or folders that have
numeric numbers rather than text descriptions that better
identify the pictures, such as “Matt’s 13th birthday party in
Hawaii.jpg.” This description of the photographs, i.e. the
metadata, is helpful in identifying the items themselves. In
addition, there are other properties that are part of metadata,
such as the date and time of the photograph that could be useful
in finding the specific photograph.
Here’s how WinFS fits into this picture. WinFS stores
information in a relational database where data is represented
as a set of attributes (fields or columns) on an item (row) of a
particular type (table). Unlike the current flat file system
that we currently use, data is organized in a structured format
which makes it easy to locate the information. The traditional
file system in use today has a flat structure. If you want to
identify a piece of data, such as a file, you have to use the
entire path to locate that file. There are several disadvantages
to the flat file system. It is cumbersome to identify data, it
is slower, and the length of the path is limited.
The tables and rows mentioned above allow convenient
manipulation of data in WinFS. In addition, WinFS provides
object-relational mappings to let users view their data as an
item. A WinFS item is the smallest unit of data that can be
manipulated by users and applications. The use of items makes
querying of data not only more logical but also more natural for
users. Imagine storing all your pictures in a database so they
can be sorted out quickly and accessed easily with a simple
query, instead of using more complicated methods available today
with today’s traditional file systems.
WinFS Architecture
If you take a closer look at the WinFS architecture you’ll
notice that it’s made up of five core components: APIs,
Services, Schemas, WinFS Core, and Data Model, as shown below in
Figure 1.
Figure 1 – WinFS architecture
APIs: The Application Programming Interface (API) is part
of the WinFX programming model and allows access to stored
files. This will help developers add query and searching
capabilities to the applications.
Services: This component includes synchronization and
rules. Synchronization will synchronize WinFS systems on your
network as well as create bi-directional synchronization
adapters that will allow you to synchronize WinFS with other
systems. The integration of services to other technologies like
Active Directory will make sharing of data between applications
much more flexible.
Schemas: This is WinFS schema for most of the information
you use, such as e-mail, appointments, multimedia, and
documents. WinFS introduces the concept of a schema definition
language to describe all WinFS types. There are predefined sets
of schemas that are included in WinFS. You can define your own
schemas or extend the existing schemas that are part of WinFS.
WinFS Core: Mainly made up of some basic operations and
file system services like security, import/export and quotas.
Data Model: Includes the basic item structure, relationships,
and the ability to extend them. More specifically it consists of
the following five concepts to describe data structures and
organizations.
• Types and subtypes.
• Properties and fields.
• Relationships.
• Constraints.
• Extensibility.
NTFS: So far all the indications are that WinFS will not
replace NTFS, at least not in the beginning. Instead, NTFS will
provide the foundation for WinFS. Unlike NTFS, WinFS will allow
you to organize and locate files rather quickly because they
will be stored with metadata that categorizes them using
user-defined properties. This WinFS storage method will be in
addition to the current method that NTFS uses to store files. If
users don’t want to take advantage of the advanced features of
WinFS, they will be able to stick with the current NTFS
functionality. The bottom line is that WinFS and NTFS will work
hand-in-hand to optimize Windows file storage system.
The Anticipated Arrival
There are going to be lots of rumors about the availability of
WinFS in coming months and years. Already people are guessing
that Microsoft will ship a WinFS beta version in late 2006.
Instead of relying on rumors or news that have not been
confirmed by Microsoft, I will only tell you what Microsoft
officials have said so far publicly. At the time of writing,
Microsoft has announced that WinFS will not be released with
Longhorn. It will be released at a future date. Microsoft is
refraining from giving any hints on what that future date might
be.
According to Windows Server Chief Bob Muglia, WinFS will also
not be a part of the server version of Longhorn that ships in
2007. "In terms of something major like WinFS, we still need to
figure it out," Muglia said. As far as Longhorn release,
Microsoft has said that it anticipates launching Longhorn
sometime in 2007. Based on Microsoft’s track record, some will
argue that it’s unlikely that the product will be shipped on
time. The only thing that’s certain so far is that Microsoft has
not given up it’s decade long dream of coming up with a storage
system like WinFS and they are working hard to eventually offer
this revolutionary file system storage on their Windows
platform.
