Hard Drive Failure Rates and Reliability
Many of our customers after a data recovery ask what we recommend to purchase to avoid a future data loss. Although there can never be a scenario in place to completely avoid data loss 100% of the time, there are some common sense measures to consider which can help. We encourage double backups with occasional triple backups depending on the data. The price of an external device is very inexpensive relative to the cost of a data recovery, it is essentially a “no brainer” to have multiple backups. That being said, I have listed what we have encountered over the years when dealing with hard drives and included some links for interesting reading which nearly corresponds with our findings over the years. From performing data recovery on over 2100 drives in the last 10 years, we have found Toshiba to be the best, with WD coming in second based on our database and anomalies associated with each manufacturer and model. Seagate drives in the last 3 to 5 years have had a horrific failure rate and even greater since the class action law suit with newly manufactured drives.
There are several enclosures populated with Seagate drives including LaCie, Porsche Design, Buffalo and others. Although the manufacturers of these enclosures are reputable, they often utilize drives as necessary to fill the supply chain without the knowledge of ongoing reliability and failure statistics. Apple Computer for example, has continued to use Seagate drives in many of their iMac computers.
We recommend Thunderbay hard drive enclosures populated with Toshiba drives which can be setup in a RAID 5 or higher configuration. They can be populated with your choice of whatever disk drive you want to use. They are extremely fast operating on Thunderbolt 2 and Thunderbolt 3. Although the software included with the Thunderbay is designed for Mac’s, the unit on its own works very well with a PC as long at the motherboard supports Thunderbolt.
As the price of NAND flash decreases, solid state drives are becoming more popular. People are under the misconception since there are no moving parts, SSD’s will be less prone to failure than a conventional hard drive. Some fundamental facts are the following:
- SSD’s uses “garbage collection” and “wear leveling algorithms” among many other techniques built into the SSD’s controller to constantly adjust for failure from bad bits.
- NAND flash has a high failure rate and a surprisingly low amount of actual read and write cycles versus conventional hard drives read and write cycles.
- Typical SSD’s controllers use hardware level encryption which can make data recovery difficult and often not possible.
Stay away from Samsung on the Solid State Side. We like Crucial SSD’s with Micron memory if used with Marvel controllers. Although this combination is the most reliable, it has slower performance compared to others. We would rather have a drive that continues to function with slower performance than a drive that is faster and prone to failure. The actual speed differences are negligible under most circumstances. Marvel has been around a long time and they are very good at what they do. They have been responsible for the controller on 99.98% of all the hard drives manufactured in the last 30 years.
In terms of general reliability, stay away from any drive larger than 3 TB unless they are set up in a RAID 5 or greater and then don’t exceed 5 TB for each drive. I mention this as a general recommendation to all my customers. As the public demand for greater storage capacity has increased, some of the manufacturing processes with an overall rush to market have yielded an inferior product due to cost constraints, general engineering, manufacturing processes and marketing promises. Whether you employ SSD’s or conventional hard drives, they all will fail at some point. Both technologies, SSD and Conventional Hard Disk Drives, have their benefits and issues. A proper backup plan should always be considered and implemented to avoid data loss.