Disk Failure Rates in Data Centres are Falling

The reliability of hard disk drives (HDDs) in data centres has been a topic of interest for many years. With the increasing demand for data storage, understanding the failure rates of HDDs is crucial for data centre operators and cloud service providers. A recent study by Backblaze, a cloud storage provider, has shed some light on this topic.

Annual Failure Rates

Backblaze’s study, which analyzed data from over 200,000 HDDs in their data centres, found that the annual failure rate (AFR) of HDDs has been declining over the years. The AFR is a measure of the number of failures per year, expressed as a percentage of the total number of drives.

Year AFR
2013 4.57
2014 3.45
2015 2.64
2016 2.22
2017 1.95
2018 1.73
2019 1.55
2020 1.37
2021 1.24
2022 1.15

As shown in the table above, the AFR has been steadily decreasing over the years, from 4.57 in 2013 to 1.15 in 2022. This decline in failure rates is a testament to the improvements in HDD technology and manufacturing processes.

Factors Contributing to Failure Rates

So, what factors contribute to the failure rates of HDDs in data centres? According to Backblaze’s study, the main factors that contribute to HDD failures are:

  • Age: The older the drive, the higher the likelihood of failure.
  • Temperature: Higher temperatures can lead to increased failure rates.
  • Vibration: Excessive vibration can cause mechanical failures in HDDs.
  • Power cycles: The more power cycles an HDD experiences, the higher the likelihood of failure.

HDD Models and Failure Rates

Backblaze’s study also analyzed the failure rates of different HDD models. The results showed that some models are more reliable than others.

HDD Model AFR
Western Digital Ultrastar DC HC620 0.64
Seagate Exos X16 0.83
Toshiba MG08ACA16TE 1.04
Western Digital Ultrastar DC HC630 1.15
Seagate IronWolf Pro 1.33

As shown in the table above, the Western Digital Ultrastar DC HC620 had the lowest AFR, while the Seagate IronWolf Pro had the highest AFR.

Conclusion

In conclusion, the failure rates of HDDs in data centres are declining, thanks to improvements in technology and manufacturing processes. Understanding the factors that contribute to failure rates, such as age, temperature, vibration, and power cycles, can help data centre operators and cloud service providers take steps to minimize failures. Additionally, choosing reliable HDD models can also help reduce failure rates.

Recommendations

Based on the findings of Backblaze’s study, we recommend the following:

  • Regular maintenance: Regularly check and maintain HDDs to prevent failures.
  • Temperature control: Keep data centres at a consistent temperature to prevent overheating.
  • Vibration reduction: Use anti-vibration mounts and other techniques to reduce vibration.
  • Power cycle management: Manage power cycles to minimize the number of times HDDs are powered on and off.
  • Choose reliable HDD models: Select HDD models with low failure rates, such as the Western Digital Ultrastar DC HC620.

By following these recommendations, data centre operators and cloud service providers can minimize HDD failures and ensure reliable data storage.