6.3.3 Resource utilization measures

What this heading covers

Consumption without a work unit is not a measure. The same product consumes more when it does more, so the useful form is a rate: memory per session, storage per record, energy per transaction. The 2023 model adds environmental sustainability under performance efficiency, and the draft gives it no separate family, which makes this family the natural home for an energy measure if a team needs one. The measure should say which resource, which instrument and which work unit.

Record position: family 5 of 40 in clause 6, under 6.3 Performance efficiency, which carries 3 families. It is one of the 40 families a programme has to map its measures to, and its elements are listed below. The heading is 6.3.3 in the public contents; the text behind it is not public.

How a measure for this family is built

Divide the resource consumed by the work performed: bytes of memory per concurrent session, gigabytes of storage per million records, kilowatt hours per thousand transactions. Name the instrument and the sampling interval. A peak and an average answer different questions, so report both where the resource can spike.

What to put in place before publication

1. Name the work unit before measuring consumption, and keep it fixed across releases.
2. Report peak and average separately, with the sampling interval.

Evidence to keep

Two questions to ask

• What unit of work is the consumption divided by?
• Was this a peak or an average, and what was the sampling interval?

Where teams usually fall short

• A memory total from one run, compared with a total from a run that did twice the work.

What the number cannot tell you

It cannot tell you whether the consumption was necessary. A product can be efficient and still waste memory on a path nobody uses.

What happens next

Capacity is the limit of the same curve. Read the two pages together if you are sizing an environment.