Performance

Engineering
Services

SPE·ED models are software-oriented. It models any type of software from novel applications to system software. It models distributed systems as well as mainframe execution environments. It models software written in any language running on any type of hardware and operating systems.

Users create and evaluate models with the following steps:

• Identify key performance scenarios and describe their processing steps
• Specify the number of software resource requests for each step (e.g., Selects, messages, screens, etc.)
• A performance specialist provides overhead specifications – the computer service requirements (CPU, I/O, etc.) for the software resource requests.
• Solve the models using analytic techniques (no contention and MVA solutions)
• View the initial assessment of performance, look for risk areas, and evaluate alternatives
• Automatically create a system execution model to represent the performance of combinations of performance scenarios executing on the computer facilities.
• SPE·ED simulates the system model environment to produce results that quantify the overall system performance.
• View visual results that indicate compliance with performance requirements and highlight potential software and hardware problem areas.
• Users easily modify software models to assess requirements, design, and implementation tradeoff decisions that affect performance.
• The SPE database archives results of studies for reporting and project tracking.

In order to automate the system model steps and enhance the visual perception of results, the computer facility models are deliberately simplified views of the system execution environment. It is not necessary to represent intricate processing details of the execution environment to study high level software tradeoffs.

Computer resource overhead specifications facilitate the analysis. This strategy for separating them from the software specification makes the models adaptable to all software, hardware, and mixed vendor environments. It also allows users to evaluate the performance of the software in various hardware environments to select the most appropriate platforms.

The overhead specifications are provided by a performance specialist, are stored in SPE·EDs database, and re-used in all models that execute in that environment.

SPE·ED simplifies model creation to let users focus on results.

For more information on how SPE·ED works, you can examine the Quick Start chapter of the User Guide (pdf).