Control-Flow Residual Analysis for Symbolic Automata
This work addresses monitoring overheads in large systems, but it is incremental as it builds on existing control-flow analysis methods.
The paper tackles the problem of runtime overhead in dynamic monitoring for system correctness by generalizing control-flow static analysis to optimize monitoring for symbolic automata, resulting in reduced monitoring instrumentation and logic, with empirical evidence from a financial transaction system.
Where full static analysis of systems fails to scale up due to system size, dynamic monitoring has been increasingly used to ensure system correctness. The downside is, however, runtime overheads which are induced by the additional monitoring code instrumented. To address this issue, various approaches have been proposed in the literature to use static analysis in order to reduce monitoring overhead. In this paper we generalise existing work which uses control-flow static analysis to optimise properties specified as automata, and prove how similar analysis can be applied to more expressive symbolic automata - enabling reduction of monitoring instrumentation in the system, and also monitoring logic. We also present empirical evidence of the effectiveness of this approach through an analysis of the effect of monitoring overheads in a financial transaction system.