The NetApp® RAID 6 equation computes the expected number of data loss events over the system mission time. The four input values include basic system parameters, hard disk drive (HDD) failure characteristics, and time distributions for RAID reconstruction and media scrubbing repair processes. The parameters are expressed by a Weibull distribution (a continuous probability distribution), which can model increasing, constant, or decreasing failure rates over time by changing the β (shape) parameter.
System Parameters

Disk model presets Unreliable SATA(1TB)
10K RPM FC (288 GB)
(Selecting another disk changes inputs)
HDD Failure Characteristics  Graph

ηp days
Time To Repair a Failed HDD  Graph

MTTROp hours
HDD Latent Sector Defects

TLd days
Media Scrubbing Characteristics  Graph

MTTS=  hours
Fix the marked parameters.
Time [Years]
Weibull Distribution
Time [Hours]
NetApp developed a closed-form RAID 6 equation that calculates the reliability of data stored on NetApp RAID-DP® technology and other, more general RAID 6 (or double-parity) RAID schemes. Compared with the original mean time to data loss (MTTDL) RAID equation that was formulated by Gibson and Patterson in 1993, the NetApp RAID 6 equation is much more realistic in both assumptions and outcomes. It considers failures that can be either operational (that is, complete disk failures) or transient, such as read errors (also called latent defects). The equation operates with nonexponential distributions for time to failure and repair that can vary over time. The equation also accounts for restorative features of modern RAID implementations, including NetApp Data ONTAP® systems such as media and RAID parity scrubbing, which proactively repair transient errors. The Web-based RAID 6 equation calculator’s preset input values for the time-to-failure and time-to-repair input distributions are based on an extensive analysis of field data collected from tens of thousands of NetApp FAS systems with hundreds of thousands of hard disk drives (HDDs) of different models, capacities, and types, including enterprise-class Fibre Channel (FC) and near-line SATA disks. These parameters can also be set to custom values to explore their effect on overall data reliability. Our goal is to provide an easy-to-use tool that can help system designers, who might not be reliability analysis experts, explore the design space of their existing next-generation RAID solutions. Our intent is for this tool to help designers understand the performance and reliability trade-offs and choose the right design with more or less aggressive proactive mechanisms, such as media and RAID scrubbing, and understand the inherent reliability of their devices, including HDDs and solid-state drives.
The RAID 6 equation computes the number of expected triple failures at time t by expressing the probability of the system’s being in degraded mode and experiencing another operational failure of the remaining D disks, expressed as the cumulative hazard rate, H(t). The probability of being in degraded mode can be due to a combination of two operational HDD failures, POO, or a combination of an operational and latent sector error. The latter can occur either as a sequence of an operational failure and a latent sector error, POL, or as a latent sector error and an operational failure, PLO. These probabilities use the mean values of the parameterized failure characteristics and repair time distributions. See the following equations:
The following table defines the quantities that are used for calculating the piece-wise probabilities that contribute to the overall probability of being in degraded mode. The cumulative hazard (or failure) rate, H(t), is computed from a Weibull distribution of the general HDD failure characteristics. In addition, the pseudo-characteristic life, ηp, is an approximation of the Weibull cumulative failure probability and is computed by equating exponential failure probability to that of Weibull at some time t.
Pseudo-characteristic life
Mean time to repair an operational HDD failure
Mean time to latent defect
Mean time to complete HDD media scrub
Cumulative hazard rate of an HDD
Relationship to MTTDL

The closed-form RAID 6 equation generally predicts a higher number of data loss events compared to with the MTTDL equation, because it the RAID 6 equation assumes non-constant failure rates with non-exponential distributions. Additionally, it accounts for data loss events due to latent sector defects. See the following equation: