Graphics Programs Reference
In-Depth Information
p active 1
p active 2
T request 1
T request 2
p requesting 1
p requesting 2
p idle
t start 1
t start 2
T watch dog
p accessing 1
p accessing 2
p acc 2 int 1
T end 1
T end 2
T end 2 start 1
Figure 3.11: Timed PN system with an example of enabling memory policy
policy is very limited, but it must be considered because of its theoretical
importance in the case of SPNs and GSPNs.
The other two policies are of greater importance from the application view-
point. They can coexist within the same timed PN system, because of the
different semantics that can be assigned to the different transitions of the
model. For a detailed discussion on this topic the reader is referred to [2, 19] .
An example with the enabling memory policy
A possible application of the enabling memory policy is provided by transi-
tions used to model time-out or watch-dog mechanisms.
When the transition modelling the time-out mechanism is enabled, it can
either fire, or be disabled by some other conflicting event that takes place
within the predefined time-out interval. If the transition modelling the time-
out is disabled because the system evolves into its next natural state, the
reading of the timer associated with the time-out transition can be discarded,
and the timer will be reset to a completely new time-out delay when the
transition becomes enabled again.
Let us consider for example the multiprocessor modelled by the timed PN
system in Fig. 3.5, augmented with the following further specification: one
of the two processors (say processor 2) has the privilege of interrupting the
other one (say processor 1) when the former has been waiting for the re-
source for an amount of time greater than a given threshold τ watch−dog .
Search WWH ::

Custom Search