Runtime Verification of Data-Centric Properties in Service Based Systems

Model-driven generation of runtime checks for system properties

“…In this paper, we focus particularly on the specification of constraints to be monitored at runtime, and tools that automatically generate runtime monitors from these specifications. In recent years, three classes of techniques have established themselves as the main focus areas for research: (1) contract-based approaches like JML [3], (2) approaches matching event traces against patterns, usually specified in a form of linear temporal logic (LTL) [8,11,12], regular expressions or context-free grammars [7,9,10,13], and extensions that include more complex data bindings and events [20][21][22][23], and (3) model-based approaches centered around UML-like models [16,17]. There is also a large body of work on other monitoring topics, such as intrusion and malware detection.…”

“…Events are user-selected method calls, and the designer of specifications can define regular expressions over these traces and develop behaviors, that is, additional code to be executed, when a match is found. Mop, its specialization JavaMop, and related techniques make use of the same concept of traces, but allow for more complex specification languages, for example, past-time temporal logic [10,21,24]. Java-MaC [25] and Par-BCL [21] go further by also monitoring program state in addition to events related to method calls.…”

“…Mop, its specialization JavaMop, and related techniques make use of the same concept of traces, but allow for more complex specification languages, for example, past-time temporal logic [10,21,24]. Java-MaC [25] and Par-BCL [21] go further by also monitoring program state in addition to events related to method calls. Kim et al [25] also recognize that it is useful to separate the description of low-level events and conditions from higher-level properties.…”