On optimizing runtime monitors, both statically and dynamically

Eric Bodden / Persyval-Lab

We will discuss Clara, a static-analysis framework for partially evaluating finite-state runtime monitors at com- pile time. Clara uses static typestate analyses to automatically convert any AspectJ monitoring aspect into a residual runtime monitor that only monitors events triggered by program locations that the analyses failed to prove safe. If the static analysis succeeds on all locations, this gives strong static guarantees. If not, the efficient residual runtime monitor is guaranteed to capture property violations at runtime. Researchers can use Clara with most runtime-monitoring tools that implement monitors through AspectJ aspects. In this tutorial, we will show how to use Clara to partially evaluate monitors generated from trace- matches and using JavaMOP. We will explain how Clara’s analyses work and which positive effect they can have on the monitor’s implementation.

In addition we will briefly discuss a recent piece of work in constructing prm4j, a novel and fully reusable Java library for efficient runtime monitoring. It is quite challenging to make such a library generic and efficient at the same time. This is why prm4j generates some specialized, efficient data structures on-the-fly, at the time the user registers the monitored property with the library. Opposed to systems such as JavaMOP or Tracematches, the specialization does not generate code, however, but rather parameterizes generic data structures directly on the heap.