Software Engineering

239
13.3 
Design correlation
The progressive assimilation through the levels of integration must be accounted for 
to establish accurate structural element performance specifications.
Computing resource–intensive processes may perform adequately when initially 
evaluated, before any integration has occurred. However, the integration detracts from 
individual structural element performance as computing resources must be shared 
with other data process threads. The intent is to understand the implications of appor-
tioning computing resources among interrelated, interdependent, and isolated design 
elements. This suggests that software execution profiles must be analyzed within the 
context of the computing system execution framework. This permits the software 
performance benchmarks to account for resource sharing, multi-user workloads, and 
other conditions that may be encountered during software product operations.
Software performance benchmarks should establish the computational dura-
tions and resource utilization requirements for the structural elements throughout 
the structural configuration. These benchmarks must extrapolate structural design 
mechanism performance to account for the integrated product’s execution profile 
constrained by the shared computing system’s resource demand and allocation strat-
egy. These performance benchmarks provide an engineering approximation of the 
performance characteristics derived from the functional architecture specifications.
The intent of this practice is to ensure that software product performance is 
designed into the structural configuration. It is not acceptable to delay focus on 
performance to software testing. By that time, the structural architectural deci-
sions have established a design configuration that impedes performance satisfac-
tion. Software performance must be an integral design consideration throughout 
the engineering of the software product. Establishing software performance bench-
marks addresses understanding how structural design elements provide construc-
tive, collaborative data processing mechanisms to satisfy performance objectives. 
Performance is realized by aggregating performance-related measures from the bot-
tom-tier design elements up through the structural configuration.