Chapter 11, Problem 11.1E

Explanation of Solution

Reliability Specifications:

Reliability testing is used to measure the reliability of the system. It is not easy to apply the reliability measurement practically. The difficulties that arise are as follows:

• Operational profile uncertainty – The operational profiles that are based on experience with other systems will not be an accurate reflection of the real use of the system.
• High costs of test data generation – Generating large volumes of data in an operational profile is very expensive.
• Statistical uncertainty when high reliability is specified – Statistically significant number of failures is generated which allow accurate reliability measurements. Few failures occur when the software is reliable and it is difficult to generate new failures.
• Recognition failure – One can identify deviations from a formal specification that they have, but if the specification is in formal language then there may be ambiguities. It is not obvious whether or not a system failure has occurred.

It is because of the above given difficulties, it is practically impossible to validate reliability specifications when these are expressed in terms of a very small number of failures over the total lifetime of a system.