A RAID MTBF calculator

raid_mtbf.py

#redundancy is the max number of survivable failures, so eg 1 for RAID5
#mtbf_array is an array of either actual mean-time-between-failures, or a nested RAID array

# RAID([100]*7,2) #7 disk RAID 6
# RAID([RAID([100]*3,1),RAID([1000]*3,1)],0) # RAID 50, 2 arrays of 3
# RAID([100,100,50,50],1) #RAID 5 with varying reliabilities

from random import random

class RAID(object):
	def __init__(self,mtbf_ary,redundancy):
		self.mtbf_ary=mtbf_ary
		self.redundancy=redundancy
		self.mtbf_cache=False
		return
		
	def mtbf(self,sims=10000,recalc=False):
		if self.mtbf_cache and not recalc: return self.mtbf_cache
		
		failtimes=[None]*sims
		failrates=[None]*len(self.mtbf_ary)
		
		for idx,item in enumerate(self.mtbf_ary):
			if type(item) is float or type(item) is int:
				failrates[idx]=1.0 / float(item)
			if isinstance(item,RAID):
				failrates[idx]=1.0 / item.mtbf(sims)
		
		for i in range(sims):
			trials=0
			failcount=0
			alive=[True]*len(self.mtbf_ary)
					
			while failcount<=self.redundancy:
				trials+=1
				for idx,item in enumerate(self.mtbf_ary):
					if random()<failrates[idx] and alive[idx]:
						failcount+=1
						alive[idx]=False
			failtimes[i]=trials
		self.mtbf_cache=float(sum(failtimes)) / len(failtimes)
		return self.mtbf_cache  #mean time to array failure