Annual data rebuild rate
The most conservative (largest) amount of data to be rebuilt on annually is:
- Annual data rebuild rate
- DB = (Cd × CU% × FRy% × d)
| Variable | Term | Definition | Example |
|---|---|---|---|
Cd |
Total Drive Capacity | Total capacity of each drive in the system. | If each drives has a capacity of 4 TB, then the Total Drive Capacity is 4
TB. |
CU% |
Percent Storage Used | Average fill ratio of the system. | If the system is only 80% full, then the Percent Storage Used would be 80% or
0.8. |
FRy% |
Annual Unexpected Failure Rate | A conservative (high) estimation of the annual failure rate of drive drives is 4%. Due to the failing drive migration feature, drive failures do not necessarily result in a loss of data on that drive; often much of the data can be saved. By monitoring drive health statistics failing drives can be identified prior to failure over 75% of the time. Therefore, less than one in four failures is unexpected and requires rebuilding for recovery. This is because data is preemptively migrated off of drives suspected to fail soon. Therefore the annual unexpected drive failure rate is only a quarter of the annual drive failure rate, or 1% if the annual failure rate is 4%. | |
d |
Drive Count | This is the total number of drives in the system. | If the system is comprised of 100 Slicestor appliances with 45 drives each, then the Drive
Count is 4,500. |
Example: Annual data rebuild rate
Assuming a drive capacity of 4 TB, fill ratio of 80%, annual unexpected failure rate
of 1% and 2,500 drives (a system with 10 PB of raw drive storage), the annual data to
rebuild is:
80 TB = (4 TB × 80% × 1% × 2500)