Disk Failures in the Real World: What Does an MTTF of 1, 000, 000 Hours Mean to You?

By Bianca Schroeder, Garth A. Gibson. USENIX Conference on File and Storage Technologies, 2007

834 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/51355a36ff42e47349c622162c9f2d97ec265a78

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Franck Cappello
2. M. Snir
3. Feng Chen
4. D. Ford
5. A. Geist
6. Phillipa Gill
7. Mohammad Shahrad
8. Jon Stearley

Citations in the corpus

• Auditing to Keep Online Storage Services Honest (2007, background): “Unfortunately, this information is hard to obtain for storage services: Only a few studies describe failures and their causes in deployed large-scale storage systems [10, 12].”
• Availability Knob: Flexible User-Defined Availability in the Cloud (2016, background): “Many studies [41, 57, 58] have shown that the time between failure in large computing systems is well modeled by the Weibull distribution.”
• Availability in Globally Distributed Storage Systems (2010, methodology): “We also develop a method for identifying which bursts are likely due to a failure domain.”
• DRAM errors in the wild: a large-scale field study (2009, background): “In production sites running large-scale systems, memory component replacements rank near the top of component replacements [20] and memory errors are one of the most common hardware problems to lead to machine crashes [19].”
• Understanding network failures in data centers: measurement, analysis, and implications (2011, background): “There also have been several studies on understanding hardware reliability in the context of cloud computing [11, 23, 24, 27].”
• To cloud or not to cloud?: musings on costs and viability (2011, unknown): “4 times lower than advertised [25].”
• An analysis of latent sector errors in disk drives (2007, background): “A recent study by Schroeder and Gibson [17] analyzed data involving the failure of about 70,000 disks over a period of .ve years and found that the failure rate increases over time.”
• A Comprehensive Study of the Past, Present, and Future of Data Deduplication (2016, background): “In contrast, in computer systems, the probability of a hard disk drive error is about 10 12 10 15 [71], [72], which is much higher than the aforementioned probability of SHA1-fingerprint collisions in data deduplication.”
• A Large-Scale Study of Failures in High-Performance Computing Systems (2006, background): “In addition to data on node outages from high-performance computing environments, we have also collected a number of failure data sets on storage failures, in particular, data on hard drive replacements in large storage systems [21].”