Difference between revisions of "Scalable Parametric Runtime Monitoring"

From FSL
Jump to: navigation, search
Line 3: Line 3:
<pub id='jin-meredith-rosu-2012-splash-submission' template='PubDefaultWithAbstractAndTitle'></pub></private>
<pub id='jin-meredith-rosu-2012-splash-submission' template='PubDefaultWithAbstractAndTitle'></pub></private>
== Technical Report ==
== Technical Report ==
<pub id='jin-meredith-rosu-2012-tr' template='PubDefaultWithAbstractAndTitle'/>
<pubbib id='jin-meredith-rosu-2012-tr' template='PubDefaultWithAbstractAndTitle'/>

Latest revision as of 21:41, 26 February 2016

[edit] Technical Report

Scalable Parametric Runtime Monitoring
Dongyun Jin and Patrick O'Neil Meredith and Grigore Rosu
Technical Report http://hdl.handle.net/2142/30757, April 2012
Abstract. Runtime monitoring is an effective means to improve the reliability of systems. In recent years, parametric monitoring, which is highly suitable for object-oriented systems, has gained significant traction. Previous work on the performance of parametric runtime monitoring has focused on the performance of monitoring only one specification at a time. A realistic system, however, has numerous properties that need to be monitored simultaneously. This paper introduces scalable techniques to improve the performance of one of the fastest parametric monitoring systems, JavaMOP, in the presence of multiple simultaneous properties, resulting in average runtime overheads that are less than the summation of the overheads of the properties run in isolation. An extensive evaluation shows that these techniques, which were derived following a thorough investigation and analysis of the current bottlenecks in JavaMOP, improve its runtime performance in the presence of multiple properties by up to two times and the memory usage by 34\%.

Personal tools