Big Parity® from StreamScale leads the industry in RAID reliability and processing speed, eliminating data loss and corruption due to disk failure, service errors, silent data corruption, and unrecoverable read errors. 30X faster than competitive RAID systems, Big Parity is the only product that provides advanced Verified Erasure Coding® which enable up to 127 parity drives to keep data safe and accurate during long data analysis cycles or rebuilds.

Service Mistakes & Delays

A storage system is crippled by drive failures. Every second the failed drive is in the storage array other errors are more difficult to detect and correct. Delays in service increase the probability that data will be lost when the number of errors exceeds what the system can correct for.

The challenge of understanding Data Loss in the real world also requires understanding human error. Whenever a human is involved in repairing a storage system there is a probability that a mistake will be made. There are examples of the person performing maintenance replacing the wrong drive, installing a corrupt drive, and damaging a drive on installation. In addition to a Service event creating data loss there is also the probability that the newly installed drive will be not be rebuilt correctly.

Simulations demonstrate that even a small probability of an imperfect repair decreases the reliability of a RAID system by several orders of magnitude. On Big Data systems with higher drive counts, the effect is even more pronounced, decreasing reliability as much as 10 orders of magnitude. To properly protect the overall reliability of a data storage system the possibility of a service mistake must be taken into account (single drive repair, multiple simultaneous drive repairs, and imperfect repair).

What does this mean in practical terms? When measuring or projecting your RAID system reliability include the probability of imperfect service and drive reconstruction problems. You must add at least one if not two or more additional check drives to the RAID array to maintain the same expected system reliability. Instead of the typical 2 parity drive RAID 6 system you would want at least 4 parity drives just to account for imperfect service.

Big Parity is a Verified Erasure Code system that can expand to 127 parity drives. Instead of just adding two parity drives for drive failure and 2 more for imperfect repair, you can use Big Parity to also add drives to account for service delays and drive reconstruction problems. This capability also has the practical application of enabling the creation of larger storage arrays.

Adding an additional check drive to a RAID array will increase the reliability of that system by an order of magnitude, whereas an additional data drive will only decrease the reliability of the system slightly. Therefore, the largest RAID systems are able to achieve the best reliability at the lowest cost per byte stored by using Big Parity with Verified Erasure Coding.

Contact us now to get more information on how your organization can take advantage of Big Parity to protect your data from Service Mistakes and Delays. Join the discussion on our Forum page and tell us about your real world story about storage reliability.