Why Software doesn't wear out explain?

Hardware Failure Rates

The illustration below depicts failure rate as a function of time for hardware. The

relationship, often called the "bathtub curve," indicates the typical failure rate of

individual components within a large batch. It shows that in say a batch of 100

products, a relatively large number will fail early on before settling down to a steady

rate. Eventually, age and wear and tear get the better of all them and failure rates rise

again near the end of the products life. To assist in quality control, many new batches

of products are 'soak' tested for maybe 24 hours in a hostile environment

(temperature/humidity/variation etc.) to pinpoint those that are likely to fail early on

in their life, this also highlights any inherent design/production weaknesses.

These early failure rates can be attributed to two things

• Poor or unrefined initial design. Correcting this, results in much lower failure

rates for successive batches of the product.

• Manufacturing defects i.e. defects in the product brought about by poor

assembly/materials etc. during production.

Both types of failure can be corrected (either by refining the design, or by replacing

broken components out in the field), which lead to the failure rate dropping to a

steady-state level for some period of time.

As time passes, however, the failure rates rise again as hardware components suffer

from the cumulative effects of dust, vibration, abuse, temperature extremes and many

other environmental maladies. Stated simply,

"…The hardware begins to wear out."Software Engineering Topic 1 Page 10

Software Failure Rates

Software is not susceptible to the same environmental problems that cause hardware

to wear out. In theory, therefore, the failure rate curve for software should take the

form shown below.

Undiscovered defects in the first engineered version of the software will cause high

failure rates early in the life of a program. However, these are corrected (hopefully

without introducing other errors) and the curve flattens as shown. The implication is

clear. Software doesn't wear out. However, it does deteriorate with maintenance as

shown below.

During its life, software will undergo changes and it is likely that some new defects

will be introduced as a result of this, causing the failure rate curve to spike as shown

above. Before the curve can return to the original steady-state failure rate (i.e. before

the new bugs have been removed), another change is requested, causing the curve to

spike again. Slowly, the minimum failure rate level begins to rise-- the software is

deteriorating due to change.

Thanks & Regards,

Bastin Vinoth NG