Click here to access our Podcast/Media archive

Listen to Gear6 Podcasts on iTunes!


  • thoughtput (at)
    Drop us a note with your thoughts or comments. Thanks!

« Building an Accelerated Archive | Main | Please Don't Move My Data »

July 25, 2007


open systems storage guy

I have a question about your alternative: when you refer to a "single, scalable, content store", are you talking about spinning disk, or more permanent media like tape or optical?

If you're talking about disk, you are not addressing the reason people use more permanent media types. Most people who have to keep permanent data need it as an archive- they need to be able to go back to a point in time and see what a certain piece of data looked like then. People who just need to separate their production data into "hot" and "warm" would be well served by simple disk with cache, but people who need archives can't afford to keep those archives on disk.

Gary O.

Great question. I am referring to a single, scalable, persistent, disk-based content store.

While archiving to tape or optical will work in cases where cost-savings trump access time to older data, others might be served by having a disk-based archive that uses snapshots for rolling back to a point in time. Of course, the single, disk-based content store still needs a suite of services than any persistent storage layer should have such as:
-backup and recovery
-and more...

I can see cases where the rate of change is such that a disk-based "accelerated archive" coupled with disk-based backup and recovery options works best. Maybe tape would be used for the most extreme recoveries. There would also be cases where the data set size and rate of change makes tape or optical a good bet for the archive, but this would typically be in cases where on-line access to the archive was less important.

The comments to this entry are closed.