Complex IT Design Proving Using SystemC 2018-03-20T16:27:49+00:00

Case Study

Design Proving a Complex IT System Specification using SystemC

This case study outlines a design proving project where the specification for a new complex IT system was modelled in order to improve the specification and build a test harness for the final system. The model implemented the full specification but on a simplified platform with stubs for the external actors interacting with the planned system. The model mocked communications with those stubs, interactions with users and the data in the system.

Enhancing the Specification

Improvements in the specification were expected to come from the following areas:

  • Removal of ambiguity: clarifications from the business analysts during implementation.
  • Demonstration of feasibility: building a working model demonstrated that the specification could be implemented.
  • System and use case testing: testing of the model enabled demonstration that the specification met the objectives of the system.
  • End user interaction: a user interface to the test harness allowed end users to interact directly with the model and give feedback on the speciation.
  • What-if analysis: by allowing different rules and values to be investigated within the model.

The Model and Test Harness

The diagram shown highlights how the test harness interacts with the model and how the users access the test harness and model. The model contains a database which is set up with initial values ‘pre-test’ via the test harness and is then queried ‘post-test’ to help establish Pass/Fail.

The Advantages of Using Accellera SystemC

The Accellera SystemC standard was used to allow modelling of multiple parallel processes within the model, interacting with multiple stubs and the test harness in parallel. The user interface allowed the user to run pre-defined tests in the test harness, or to write new tests, and to analyse the results. In this way the test harness allowed business analysts and end users to fully investigate the specification via the model. The model had a set of variables and rules (the ‘configuration’ in the diagram) configurable from the user interface that allowed the business analysts to perform a number of what-if analyses to improve the specification.

The test harness is also used to generate a set of user acceptance tests for the final system. Working with the business analysts, a set of use case scenarios are generated which are necessary and sufficient to ensure a correct implementation. The test harness can then generate the expected outputs and end state for the system for each test in turn, thus forming a full test specification.

Finally, it is worth noting that the test harness can potentially be used in two additional ways:

  • As a test oracle for the eventual full implementation of the specified system
  • As an integration test for the external actors by replacing the stub with the real actor.

Project Deliverables / Benefits

The project delivered the following:

  • A working, executable model of the complete specification.
  • A test harness which demonstrated that the specification fully met the required use cases.
  • A user interface that allowed the end users and business analysts to run pre-defined tests, run their own tests and perform what-if analyses.
  • A test specification defining inputs and expected outputs.

The Results

The modelling activity identified a large number of holes and ambiguities in the specification and the what-if analyses allowed a number1 of specification improvements to be made. This should result in both time and cost savings in the actual implementation phase that easily justify the modelling costs.
The test harness is expected to be reused with the actual implementation and to enable independent integration testing for external actors.

Note 1: The exact numbers are commercially confidential.

Get in Touch

Find Out More

Contact one of our consultants today to discuss your requirements.
No hard sales, just pertinent questions to understand your needs and to discuss how we may be able to help.

Alternatively contact one of our Local Sales Offices.

Get in Touch
T&VS NEWSLETTER SIGN-UP
The T&VS newsletters inform you about industry news, events and information from T&VS. No spam, we promise and it is always easy to unsubscribe.
We never share your information. Read our Privacy Statement
Interested in Formal Verification?
Then why not attend the TVS Formal
Verification Bootcamp training?
The 2-day Formal Verification Bootcamp is for design and verification engineers looking to enhance their knowledge of formal verification and to learn how to write effective assertions to find and fix bugs. The course is a mix of presentations and hands-on development exercises.
Bootcamp Enquiry Form
If you are interested in receiving additional information on the course then simply email Mike Bartley (TVS CEO and Course Leader) by entering your details below.
Interested in SystemC?
FREE SystemC UVM Library Now Available
The TVS SystemC UVM library closely mimics UVM but gives users a license free UVM-based verification environment.
Have your product requirements been successfully tested and implemented?
Find out how asureSIGN can help you implement a successful Requirements Driven Verification and Test Strategy by visiting asureSIGN or enter your details and we will be in touch.
Course Dates and Pricing
To receive additional information, including course dates and pricing, please contact our training team who will be happy to help.
Download Request
Please complete the following form then click 'submit' to access the download.
Presentation Request
Please complete the following form then click 'submit' to gain access to the presentations.
DOWNLOAD REQUEST
Please complete the following form and then click 'submit' to gain access to the download.
FREE QA ASSESSMENTS
Did you get what you were looking?

Let the testing experts help. We will run a FREE QA assessment which will include our top 5 recommendations to help maximise your testing.