Seed of a Great Idea

In the darkrecesses of my brain is the memory of Microsoft’s FrontPage and itssubstitution of a Web server for a traditional file system. In several, olderreleases of FrontPage, the Web server was mandatory and provided the platformfor file and content storage. That led to problems for developers who did nothave access to a reliable Web server. So Microsoft, in FrontPage 2000, hasrelented and les Web developers use their native file system as the foundationfor Web development.

But using aWeb server as a platform for storage and collaboration was the seed of a greatidea: Why not combine the benefits of a traditional file storage system and onebased on a Web server? The idea has blossomed into a new tool that Microsoftcalls the Web Storage System.

I’ve had achance to try the recent Release Candidate 2. While I think it has promise, itdoesn’t seem like a new idea as much as it seems like an old idea warmed up andsent toward a new target.

The firstthing to know is that the Web Storage System is built on top of ExchangeServer. Those who use Windows 2000 for Internet infrastructure but use othermessaging and collaboration systems are likely to be disappointed that the WebStorage System requires Exchange.

Thinkingabout this -- and trying out Microsoft’s sample applications -- I’ve come tothe conclusion that Web Storage System was intended to be tightly integratedinto higher-level collaboration and messaging tools. You can access items inthe Web Storage System through Active Server Pages and other applications thatuse Collaboration Data Objects for Exchange Server as well as the venerableMessaging API. This means Web-based access to data can be optimized and thatapplications that leverage Internet technologies and storage can flourish, suchas Instant Messaging applications that support transfer of images or reports inreal time or Internet-based time management applications.

The dataitself is hosted as containers using Exchange’s native storage engine. You canaccess any kind of data in those containers in three ways: using the Web’s HTTPprotocol with WebDAV extensions, using Exchange’s OLEDB provider, and using thestandard file system.

The WebStorage System looks, at first, like any hierarchical file system: folders,subfolders, and objects stored in the folders. On the face of it, that doesn’tadd much value. But consider this: What if you could associate securitydescriptors and information schemas with all the information you stored?

The WebStorage System makes it possible to build schemas for data and other objects inthe store. This isn’t a new idea, but it does make possible the ability to useHTTP to find out the properties of objects in storage before they areretrieved, replaced, or changed. Part of the schema could be a specific displayformat in HTML and Cascading Style Sheets stored with the objects themselves.One feature that the Web Storage System includes is a template system thatmakes it possible to hide the differences in how different users’ softwaredisplay a particular object.

Folders andobjects can also have access control schemas built using XML-formatted securitydescriptors. That makes it possible to develop client-platform-neutralapplications that take advantage of the Web Storage System. In fact, thoseschema can simply be a “property” stored with each folder or object. Any numberof properties -- including security and access control -- can be stored with eachitem, and the set of properties for each item can be different from any otheritem.

As I triedRelease Candidate 2 of Web Storage System I was stuck by a strong sense of déjàvu. First, the idea of a Web-based file system -- like the idea in early versionsof Front Page -- clearly has enough merit to be tried a second time. Second, asimple replacement for the elderly native file system doesn’t seem to be enoughfor applications vendors.

Vendorsrealize that databases bring sophistication and function to traditional storagetasks. Oracle’s IFS is an example of a vendor rejecting simple file systems infavor of databases that act as the file system and add value to it. Microsoft’sWeb Storage System is another example of this trend. No doubt we’ll soon see thesame strategy from other vendors, such as Novell and Sun Microsystems.

Will it payoff? Despite the fact that its middle name is “storage,” Microsoft’s new toollooks more like a collaboration application enabler than a radical new approachto storage. Tied tightly to Exchange, I doubt Web Storage System will be ableto spread its wings and take off as an independent storage tool for systemadministrators. Tied tightly to Oracle, I doubt IFS will ever have the chanceto achieve the market share needed to become more than a niche solution.

That’s toobad. As I said, there’s a seed of a great idea here. --Mark McFadden is a consultant and is communications director for theCommercial Internet eXchange (Washington). Contact him at mcfadden@cix.org.