This paper gives a brief overview of Monitoring Oriented Programming (MOP). In MOP, runtime monitoring is supported and encouraged as a fundamental principle for building reliable software: monitors are automatically synthesized from specied properties and integrated into the original system to check its dynamic behaviors. When a specication is violated or validated at runtime, user-dened actions will be triggered, which can be any code from information logging to runtime recovery. Two instances of MOP are introduced: JavaMOP (for Java programs) and BusMOP (for monitoring PCI bus trac). The architecture of MOP is discussed, and a brief explanation of parametric trace monitoring and its implementation is given. Finally, a comprehensive evaluation of JavaMOP attests to its eciency, especially with respect to similar systems; BusMOP, in general, imposes 0 runtime overhead on the system it is monitoring.
分享
引用
