Objective
Due to growing complexity of software in telecom applications exhaustive testing becomes too expensive and is in contradiction with Alcatel's goal to reduce time to market by 30%. Testing of large systems generally suffers from this coverage versus cost dilemma. This may lead to late delivery and insufficient quality. On the other hand software process improvement leads to an improved quality of software before test. Therefore an exhaustive testing isn't efficient. Nevertheless testing is still necessary, but the focus of testing moves from error detection to certification of quality.
With the introduction of statistical testing based on a usage specification we want to improve the reliability of our software by 20% and be able to predict it. As a consequence we hope to accelerate the acceptance test by 20% and reduce the maintenance effort by 20%. The efficiency of tests shall be better than conventional testing.
THE EXPERIMENT
In order to certify and to improve the reliability of software by testing, it is necessary to regard the customers' view on reliability. The experiment will introduce usage modelling and statistical testing for a subsystem of a quite large customer project, and will involve the customer in the early phases. Customers count failures per execution, whereas developers count defects per kilostatement. Statistical testing means random selection of representative test cases from the infinite number of possible executions of the software. The expected usage of the system is modelled with Markov-Chains. Test cases are generated from the usage model and run on test equipment. The test results are evaluated in order to predict the reliability of the system. The reliability of the delivered system is monitored and compared with the prediction. The modelling, test generation and evaluation is carried out using toolSET_Certify.
EXPECTED IMPACT AND EXPERIENCE
Alcatel SEL expects, that statistical testing will improve the reliability of the delivered system and that this reliability can be predicted. The benefits of increased reliability will be accelerated customer acceptance test and reduced maintenance effort. We expect also an increased efficiency of tests.
Fields of science (EuroSciVoc)
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
You need to log in or register to use this function
Call for proposal
Data not availableFunding Scheme
ACM - Preparatory, accompanying and support measuresCoordinator
75175 Pforzheim
Germany