Test Isolation Policy for Safe Runtime Validation of Evolvable Software Systems

Abstract

Runtime testing of dynamically evolvable systems is known as a complex and challenging activity. Undesired side effects may be introduced as the interference between test cases and business functionalities of the components under test. To handle such problem, test isolation strategies are used to execute runtime tests safely. In this proposal, we discuss the need to adapt the well known test isolation mechanisms to cope with heterogeneity of the components under test. For doing this, we define a Test Isolation Policy that produces the suitable test isolation solution depending on the testability degree of each component under test. We show its usefulness by the means of a case study in the healthcare domain, called Teleservices and Remote Medical Care System (TRMCS). Some experiments are carried out to show the efficiency of our approach and the low overhead it introduces.