Information Lifecycle Management (ILM) ideas have data moved between different tiers of storage so as to reduce its storage cost. Instantly-needed data goes on fastor primary disk. Less quickly-accessed data, possibly termed reference data, goes on slower and more capacious secondary or nearline disk storage. Archive data goes onto tape. These are the three basic ILM tiers. One file system covers both primary and secondary disk. Generally an archive structure is used on tape because the blessed medium cannot support random I/O.

Backup is a form of ILM, in that it transfers data to a tape tier of storage, but isn't in existence for ILM reasons. It's there because disk is unreliable and users delete files by mistake. Again, because tape can't support random I/O, a disk-based file system isn't used and data is contained in backup sets.

Replication and snapshots are fancy ways of 'backing up' data from disk to disk such that it is instantly usable if the primary data is unavailable. Disk-based backup uses disk as the target for backup. Because of this it doesn't have to use backup data sets - disk, obviously, can support disk-based file systems. However, for convenience of interfacing to existing backup applications, the disk-to-disk backup applications typically emulate tape.

Can ILM and data protection technologies be combined in one unified product?

Logically they could. ILM needs a data moving engine and a backup application, or an archive application, is simply a specialised data moving engine. However, it would be radically easier to combine ILM and data protection if the same file system could be used on both disk and 'back up/archive' storage tiers.

That is going to require random access media for the backup and archive tier and that means, in turn, no tape. Any random access media for backup and archive applications can't stream; it has to spin. That means that optical disks are the only viable - I should say remotely viable -media for this and, further, holographic disk is the only optical technology with the capacity required and the other must-have features of media removability and write once; read many (WORM) technology.

In practise I just can't see this happening. I could see an ILM product include the integration of backup applications so as to bring in tape media as a tier the ILM faclity can handle. But how it would actually happen I just don't know. Perhaps EMC, since it has a strong ILM focus plus Legato and Dantz backup products, might get a jump on the other vendors. Sun's Data Management Group, StorageTek that was, has a great set of tape products and a terrific set of ILM ideas but no backup software. It might come up with an ILM tool that contains a place for backup applications to be plugged in.

Until these things happen ILM is going to be crippled because it doesn't really manage tape as a tier in its storage tiers. ILM is, today, in my view, fundamentally about disk. That's my spin on the subject.