pub

Implementation check, performance, monitoring and representational twin alignment.

Monitoring and representational twin

Checks compare the implementation with a representational design (digital twin or model). They are one of three types. Benchmark (quantitative analysis) provides specific metrics which are comparable across platforms, software versions and datasets. Bencharks provide overall performance data and include generic CPU, Network and Memory USE (Utilization, Saturation and Error) monitoring. Unit (functional analysis) testing insures components are orchestrated according to the design and interacting correctly. Profiling (qualitative analysis) of implementation, a break out of components and system calls into relative ratios of time and resource consumption. Profiling incurs a high overhead, skews benchmark metrics, and only used as an analytic tool.

Identify how closely the implementation twins the Design plan.

Benchmark - quantitative analysis

High level checks, basic operation checks and performance reporting.

Unit - functional analysis

Compare the implementation primitive Registration components with the Functional Plan.

Profiling - qualitative analysis

General development resources, diagnostic software, analytic tools and component tutorials