Show simple item record

dc.contributor.authorJoshi, Yogi 13:20:31 (GMT) 13:20:31 (GMT)
dc.description.abstractRuntime verification (RV) is an effective and automated method for specification based offline testing as well as online monitoring of complex real-world systems. Firstly, a software framework for RV needs to exhibit certain design features to support usability, modifiability and efficiency. While usability and modifiability are important for providing support for expressive logical formalisms, efficiency is required to reduce the extra overhead at run time. Secondly, most existing techniques assume the existence of a complete execution trace for RV. However, real-world systems often produce incomplete execution traces due to reasons such as network issues, logging failures, etc. A few verification techniques have recently emerged for performing verification of incomplete execution traces. While some of these techniques sacrifice soundness, others are too restrictive in their tolerance for incompleteness. For addressing the first problem, we introduce RitHM, a comprehensive framework, which enables development and integration of efficient verification techniques. RitHM's design takes into account various state-of-the-art techniques that are developed to optimize RV w.r.t. the efficiency of monitors and expressivity of logical formalisms. RitHM's design supports modifiability by allowing a reuse of efficient monitoring algorithms in the form of plugins, which can utilize heterogeneous back-ends. RitHM also supports extensions of logical formalisms through logic plugins. It also facilitates the interoperability between implementations of monitoring algorithms, and this feature allows utilizing different efficient algorithms for monitoring different sub-parts of a specification. We evaluate RitHM's architecture and architectures of a few more tools using architecture trade-off analysis (ATAM) method. We also report empirical results, where RitHM is used for monitoring real-world systems. The results underscore the importance of various design features of RitHM. For addressing the second problem, we identify a fragment of LTL specifications, which can be soundly monitored in the presence of transient loss events in an execution trace. We present an offline algorithm, which identifies whether an LTL formula is monitorable in a presence of a transient loss of events and constructs a loss-tolerant monitor depending upon the monitorability of the formula. Our experimental results demonstrate that our method increases the applicability of RV for monitoring various real-world applications, which produce lossy traces. The extra overhead caused by our constructed monitors is minimal as demonstrated by application of our method on commonly used patterns of LTL formulas.en
dc.publisherUniversity of Waterloo
dc.subjectruntime verificationen
dc.subjectlossy tracesen
dc.subjectruntime verification frameworken
dc.titleRitHM: A Modular Software Framework for Runtime Monitoring Supporting Complete and Lossy Tracesen
dc.typeMaster Thesisen
dc.subject.programComputer Scienceen Science (David R. Cheriton School of)en
uws-etd.degreeMaster of Mathematicsen

Files in this item


This item appears in the following Collection(s)

Show simple item record


University of Waterloo Library
200 University Avenue West
Waterloo, Ontario, Canada N2L 3G1
519 888 4883

All items in UWSpace are protected by copyright, with all rights reserved.

DSpace software

Service outages