This is the conclusive diagram presented in Storage Research conference SYSTOR10 hosted by IBM.
It shows the effectiveness of defragmentation as detected in different states using Lace level units.
The units on both axis are called Lace Level and they reflect the disorder level on the logical disk. The calculation is best described in the above TECHNOLOGY page.
Briefly - they show the difference between the measured disk - against a theoretical disk having the same size and contain the same data with the files perfectly located to allow best results for read and write operations.
The numbers are based upon the research covered more than 450,000 real disks from all around the world with different capacity and different patterns.
In order to produce the above diagram, the results of all the disks that was measured "Before and after" defragmentation was analyzed, and was grouped according to the initial fragmentation level.
The y-axis shows the improvement of the defragmentation (The difference between the initial and final fragmentation)
You can see that:
When defragmentation was performed on disks with a high fragmentation level (The RED columns), the efficiency was poor. Those disks - after being identified by our tools - are target to further maintenance (i.e. replacing disk, replacing computer etc...)
With disks with a low fragmentation level (GREEN), the results are poor too, and it proves that it was a waste of time and an extra risk to run the defragmentation process.
Defragmentation of the disks in the middle (YELLOW) yield the best results.
The main benefit of our products are for the yellow disks - but you need to monitor all of them.
