About S.M.A.R.T.
Ground rules for SMART industry projects are:
- Work to date has been separate, cooperative research projects, with each company controlling their own project. Drive customers have not yet been involved in these projects.
- A typical successful project resulting in four-times higher failure warning accuracy at low false alarm rates is described in Improved Disk-Drive Failure Warnings, IEE Transactions on Reliability.
- We offer to assist and advise on drive company validation testing of our improved SMART (such testing is required to set the warning trigger threshold, as with today's multiple-threshold SMART).
- Drive manufacturers identity in data is held confidential.
- Study does not affect the manufacturers terms of sale or warranty of drives or storage arrays.
- We are not asking for funding from the companies.
- Although SMART involves industrially sensitive technology and reliability factors, the semiconductor industry has succeeded in similar studies, such as the well-known "Benchmarking Semiconductor Manufacturing" study (IEEE Trans. Semiconductor Manufacturing, vol. 9, no. 2 158-169, May 1996). Wafer fab plant yield data was not identified by company, yet each participating company could identify its own data for comparison.
- A example of a drive reliability study (pre-SMART), published with drive manufacturer anonymity: Ting-Ting Y. Lin (UCSD) and Daniel P. Siewiorek (CMU) "Error Log Analysis: Statistical Modeling and Heuristic Trend Analysis", IEEE Transactions on Reliability, Vol. 39, No. 4, October 1990, p.419-432.
- Future meetings between the companies may prove beneficial; if each company decides to participate.
