One implementation of the CPU-cache-main memory subsystem of the PDP 11/70 consists of the following chip types :
Chip type | Number of chips, nᵢ | Failure rate per chip (number of failures/10⁶ hours), λᵢ
SSI | 1,202 | 0·1218
MSI | 668 | 0·242
ROM | 58 | 0·156
RAM | 414 | 0·691
MOS | 256 | 1·0602
BIP | 2,086 | 0·1588
It is assumed that the times to failure of all chip types are exponentially distributed with the failure rates (λᵢ) shown above. All the chips must be fault-free in order for the system as a whole to be fault-free.
(I) Find the system time-to-failure.
(II) Find the mean lifetime of the system.
(III) What is the probability that the system will still function after 1,000 hours of use ?
10 marks · directive: Find
The core of a good answer
The chips form a series system, so the system time to failure is exponential with rate Λ = Σ nᵢλᵢ = 1,205·85 failures per 10⁶ hours = 0·0012058 per hour, i.e. R(t) = exp(−0·0012058 t); the mean lifetime is 1/Λ ≈ 829·3 hours, and R(1,000) = e^(−1·2058) ≈ 0·299.
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
A series reliability problem with exponential components: the minimum of independent exponentials is exponential with the sum of the rates. Compute each nᵢλᵢ, add, convert units from per 10⁶ hours to per hour, then give the distribution, the mean and the survival probability at 1,000 hours.
Answer outline
Introduction
Assume chips fail independently. The system works only if all N = 4,684 chips work, so T_sys = min over all chips of Tⱼ, and R_sys(t) = Πⱼ Rⱼ(t) = Πᵢ exp(−nᵢλᵢt).
Step 1: Contribution of each chip type (failures per 10⁶ h)
- SSI: 1,202 × 0·1218 = 146·4036
- MSI: 668 × 0·242 = 161·6560
- ROM: 58 × 0·156 = 9·0480
- RAM: 414 × 0·691 = 286·0740
- MOS: 256 × 1·0602 = 271·4112
- BIP: 2,086 × 0·1588 = 331·2568
- Λ = Σ nᵢλᵢ = 1,205·8496 failures per 10⁶ hours = 1·2058 × 10⁻³ per hour.
Step 2: (I) System time to failure
- R_sys(t) = exp(−Σ nᵢλᵢ t) = exp(−0·0012058 t), t in hours.
- So T_sys ~ Exponential(Λ = 0·0012058 per hour); density f(t) = Λe^(−Λt), constant hazard Λ.
Step 3: (II) Mean lifetime
- MTTF = 1/Λ = 10⁶/1,205·85 = 829·29 hours (about 34·6 days of continuous operation).
Step 4: (III) Reliability at 1,000 hours
- R(1,000) = exp(−0·0012058 × 1,000) = exp(−1·2058) = 0·2994.
- About a 30% chance the subsystem is still fault-free after 1,000 hours.
Conclusion
T_sys is exponential with rate 1·2058 × 10⁻³ per hour, mean life about 829 hours and R(1,000 h) ≈ 0·30. Check: 1,000 hours exceeds the MTTF, so the reliability must be below e⁻¹ = 0·368, which it is. RAM, BIP and MOS chips together contribute about 74% of the failure rate, so they are where redundancy or better parts would help most.
Examples and data to cite
- MIL-HDBK-217F, Reliability Prediction of Electronic Equipment, provides failure rate data for military electronic components; for a series system of independently failing parts the total failure rate is the sum of the parts' rates.
Facts to remember
- Series system: R_sys(t) = Π Rᵢ(t).
- Minimum of independent exponentials with rates λᵢ is exponential with rate Σλᵢ.
- For exponential life, MTTF = 1/λ and R(t) = e^(−λt).
- Failure rate units: failures per 10⁶ hours must be divided by 10⁶ to get per hour.