Data centers have recently reported an increased rate of hard drive failures. Some brands are failing at a much higher rate than expected, prompting questions about the suitability of hard drives for use in data centers.
Online data backup company Backblaze recently completed a study that looked into failure rates for different brands of hard drives. This study found that annual failure rates for Seagate are much higher than those for some other brands. In particular, the rates of failure for 3.0 TB Seagate hard drives increased significantly between January and September 2014. Failure rates for 3.0 TB Western Digital hard drives also increased during this time.
Data centers expect a certain number of hard drive failures to occur each year. However, increased rates of hard drive failure increase costs for these centers, which need to store huge volumes of customer data. As a result, companies such as Backblaze collect data about the reliability of each brand they use. Data centers need to have a consistent capacity so they can meet the growing demand for cloud storage.
Not all hard drives are failing at an increased rate. Backblaze data shows that Hitachi hard drives, as well as 4.0 TB Seagate drives and 1.0 TB Western Digital drives, are performing well. Hitachi hard drives have annual failure rates of only around 1 percent, compared to 15 percent for Seagate 3.0 TB drives, an increase from 9 percent in the previous year.
This study raises the question of why certain drives have started to fail more often recently. According to Backblaze, the reason for particular hard drives failing at much higher rates than others could be that the less reliable drives are not suited to the data center environment. This environment places particular demands on hard drives, which are very different to those experienced in a home or office environment. As a result, the results of studies conducted in data centers might not reflect how those brands perform in consumer environments.
Another possible reason for the high failure rates is that many of Backblaze's data center hard drives were obtained through a process known as drive farming. In order to meet their data storage needs in the wake of a global shortfall in supply, Backblaze bought up hard drives where they could. Rather than buying the internal drives required for the data center, the company bought external drives, which were more easily available at the time. These drives had to be removed from their casings, which could explain their enhanced failure rates.
Drive farming is one possible reason for increased hard drive failure rates. However, some hard drives may simply not be able to perform in the demanding environment of the data center. By collecting data about hard drive failure rates, online backup and cloud storage companies ensure they consistently meet their customers' needs.
Photo courtesy of BrandonSigma at FreeDigitalPhotos.net