One of the most requested features we have on this site is for a RAID reliability calculator. Over the past few years, one of the big questions is how can a user get a rough ballpark estimate about which RAID levels are the most risky to implement. About two years ago I published the RAID Reliability Anthology primer which gave a good general primer on what impacts reliability. I used a simple Mean Time to Data Loss model, or MTTDL for short, that simply looks at drive reliability. After publishing that piece, I had several folks that work at major data storage firms tell me that the model was too simple, it did not take into account things such as motherboards failing, RAID controllers failing, ECC failures and etc. The other big point was that users wanted to be able to plug in their own numbers. Today we have an answer to at least one of those feedback points. That is really exciting!
After speaking with a few industry experts, I came to the conclusion that the model is not accurate if one were trying to publish in the IEEE. At the same time, it did provide a simple model that is a quick and dirty web version that folks can use.
My next challenge was the realization that I needed to turn the RAID reliability calculator spreadsheet into an online calculator for STH’s readers to use. Sadly, while I am very comfortable in Excel, I am less comfortable in PHP and HTML5. After two attempts to get this developed, a promising programmer in the Silicon Valley, Simone Willett, came to our aid. Her efforts involved taking a fairly complex Excel model and turning it into a web version. What we are launching today was a considerable effort on her part. Simone is also one of the big drivers behind the RAID Calculator that was released earlier this year, as well as its subsequent refinements. A big thank you to her for continued efforts.
If you want to see the basics of how this model works the Reliability Anthology is probably a good read.
Getting Started with the RAID Reliability Calculator
If you want to play with the RAID Reliability calculator, you can check it out on the official RAID Reliability Calculator page (yes it needed its own page because it did not fit in the standard site theme.) Our ask is that if you do have feedback, please provide it in the forums. The goal is to make the best tool available for folks and an important part of that is receiving and taking action on feedback.
More to come as this is one of the long term visions for the site. If you do have programming expertise and want to help on one of the future efforts, please feel free to get in touch with me.