Since the fixed drive allocations limit what's available to any given user or group, Intel helps speed things along with mechanisms to update and manage stored drive images. If enough room to store a new backup isn't available, the oldest incremental backup data is automatically merged into the original backup data set. Of course, this also means that older incremental backups can no longer be accessed individually.

In the event of a total data loss, such as when the partition table on a drive becomes unreadable, it's natural for users to want to restore their backups as quickly and automatically as possible. Alas, this is where the user interface falls short. Because the software must be able to work from "bare iron" in a worst-case scenario, recovery is done using a bootable Linux CD.

Unfortunately, the boot CD's capabilities are minimal, and require using command-line input. You need to supply all relevant computer information, including computer name, server name, etc. and select the backup you want to restore from a list of available backup sets. This whole process creates plenty of opportunities for novices to make mistakes. Again, according to Intel, they expect to improve this process in a future software release.

Of course, doing a large file restore across even a gigabit network is going to be slower than from a directly-attached drive. It took more than two hours to restore a 45 GB sample partition, even though that partition contained only 12 GB of actual files. Since current computers come equipped with 100+ GB drives, it could take days to restore a small workgroup worth of machines from the SS4000-E