A flexible simulation tool for estimating data loss risks in storage arrays

Proteus is an open-source simulation program that can predict the risk of data loss in many disk array configurations, among which, mirrored disks, all levels of RAID arrays and various two-dimensional RAID arrays. It characterizes each array by five numbers, namely, the size n of the array, the number nf of simultaneous disk failures the array will always tolerate without data loss, and the respective fractions f
₁
, f
₂
and f
₃
of simultaneous failures of n
_f
+ 1, n
_f
+ 2 and n
_f
+ 3 disks that will not result in a data loss. As with any simulation tool, Proteus imposes no restriction on the distributions of failure and repair events. Our measurements have shown a surprisingly good agreement with the results obtained through analytical techniques and no measurable difference between values obtained assuming deterministic repair times and those assuming exponential repair times.