BridgeHead's product range is firmly rooted in the storage and media management markets, and HTFileStore is no exception. It's a file archiving system that allows you to define and implement file archiving policies on networked filestores.

The idea of file archiving is a simple one. Most organisations have some kind of issue with their main server disks filling up with stuff that isn't used very often. It makes sense, therefore, to have some kind of process that lets you keep the important stuff on line whilst farming off older stuff onto backing store. Although small organisations can write stuff manually to tape or CD, this quickly becomes unmanageable and an automated archive/recover process is required. This is where HTFileStore comes in.

The application runs on Windows servers (W2000 and later), and is managed via a Java (and thus cross-platform) console applet that looks rather like Microsoft's MMC. There are three key classes of "node" in an installation: the "control nodes" (the servers that do the job scheduling and execution), the "service nodes" (the things whose files the archiving process works on) and the "backup nodes" (the servers to which the backing store is connected).

Server volumes are accessed via standard Windows sharing protocols, and the backing store is managed by OpenMedia, BridgeHead's long-established media management suite. Although the actual structure of the backing store is as complicated as you want to make it, once you've set up OpenMedia you simply refer to the repositories as simple objects in the various GUI tools.

Your archiving regime is defined by creating "policies" via the GUI. A policy is basically a description of the files to work on (file types, ages, last-modified dates, locations on disk, and so on), and an action to take on files that match the criteria. Things you can do with files range from a simple "log its existence" (useful when you're doing your initial scans to see what kind of stuff is cluttering up your world), through "archive and keep a stub" (which archives each file and uses Windows' "offline" flags to make it look like it's still there) and "archive" (move to backing store and remove all traces from the original disk), to "delete" (blow it away because we don't want Megadeth MP3s on our server). Policies can be scheduled via the GUI, as you'd expect.

The main benefit of an automated archiving system over a manual approach is when it comes to retrieving files. Because it's heavily oriented toward Windows' own archiving concepts (not least the "offline" attribute that's understood by Windows 2000 and later) it's Windows desktops that are best served for retrieving files. A small plug-in integrates the Windows desktop with the archive (you get a little "My Archive Places" icon once it's installed) which means that you can see, via annotations to icons, which stuff is on-line and what's been archived. When you double-click a file that's archived, it automatically catches the open-file request and asks for the file from backing store (though because Windows controls most of the process, you mightn't get a lot of progress information when it's gone off to search a tape). There's also the option of explicitly using the taskbar-based retrieval widget, which is rather more informative when it comes to progress info. For those who don't have Windows on the desktop (eg. Mac users whose networked datastore lives on Windows and is managed by HTFileStore) there's also a web-based archive search/retrieval tool.

An important aspect of archiving is reporting – you need to know what's going on in the archive, and of course the initial setup process will involve a great deal of digging for information to allow you to choose the best archive policy. The reporting engine is very versatile, and its wizard-based nature makes it reasonably straightforward to construct reports, but its usability is the main disappointment of the product suite. The report construction process is a bit unfriendly (i.e. instead of being able to choose from a list of file types such as "Word processor documents" or "Music files" you have to type lists of file extensions). Also, it would be heavenly to be able to click on a column in a summary report and be taken to a drill-down detail report (like NetScout’s, for instance) but sadly the reports, although they're excellently laid out, are merely static documents. To their credit, the developers seemed genuinely interested when we spoke to them about how they might make the reporting side nicer to use.

All in all, HTFileStore is a very nice archiving tool. It works independently of your applications (it just looks at the filestores, after all) and it makes good use of Windows' offline storage concepts in its server and desktop integration. The reporting niggles we've already mentioned are just that – niggles – and not show-stopping problems. Although the access control layer of HTFileStore could do with being integrated into your directory service, the developers are actively considering doing so; it hasn't been done already simply because although HTFileStore is Windows-only at present, underlying components such as the Media Manager work with a shedload of non-Windows platforms, so stuff like this has to be done in the context of the whole product, not just as a Windows-specific kludge.


Automated archiving tools are the only sensible option for all but the most trivial archiving operation, as manual archiving rapidly becomes messy and impossible to manage sensibly