Information Technology Reference
In-Depth Information
system configurations. We conduct the experiment by showing how the verifica-
tion responds with regard to attributes that might have an effect on runtime. These
attributes include:
•
Components and properties: how the verification scales with regard to number of
components and properties, including also the presence of the alternative proper-
ties.
•
Nature of properties: different properties may require different expressions in the
CSP, including operations on fixed values, intervals, or more advanced operations
such as ones used to define resource constraints (e.g. sum, min, etc.).
Component-based System M
sys
(System Under Test)
...
M
M
M
...
M
M
M
Environment Component M
E
Fig. 5
System configuration used to conduct the experiments (
M
- component,
M
E
- environ-
ment component)
Figure 5 shows the system configuration used to conduct the experiments. The
inputs for the verification are provided by the Environment component, which en-
closes the component-based system under test. All experiments were executed on
Intel i7-3630QM, 4 cores, 2.40GHz.
4.1
Quantitative Results
For this experiment, we performed two measurements. In the first measurement,
we show the response time with regard to the number of components, properties
and alternative properties, having specified assumptions and guarantees as intervals.
Then, in the second measurement, we use the same configurations but with fixed
values for expressions. With these two measurements, we are able to observe the
limits on modeling the component-based system with regard to number components,
properties, and expressions used to describe the properties.
4.1.1
Measurements
In the first measurement, we execute several thousands of system configurations
with the varying number of components and properties. The measurement has two
parts. In the first part, we verify the system configurations with the varying number
Search WWH ::
Custom Search