Hardware Reference
In-Depth Information
According to this method, the load is computed in all intervals from the current time
t
and each deadline (
d
i
) of the active jobs. Hence, the intervals that need to be con-
sidered for the computation are [
t, d
1
], [
t, d
2
], ..., [
t, d
n
]. In each interval [
t, d
i
], the
partial load
ρ
i
(
t
) due to the first
i
jobs is
ρ
i
(
t
)=
d
k
≤d
i
c
k
(
t
)
(
d
i
−
,
(9.2)
t
)
where
c
k
(
t
) refers to the remaining execution time of job
J
k
with deadline less than
or equal to
d
i
. Hence, the total load at time
t
is
ρ
(
t
)=max
i
ρ
i
(
t
)
.
(9.3)
Figure 9.1 shows an example of load calculation, at time
t
=3, for a set of three
real-time jobs. Then, Figure 9.2 shows how the load varies as a function of time for
the same set of jobs.
ρ
1
(
t
)=2
/
3
J
1
ρ
2
(
t
)=3
/
4
J
2
ρ
3
(
t
)=4
/
6
J
3
0
1
2
3
4
5
6
7
8
9
10
ρ
(
t
)=3
/
4
t
Figure 9.1
Instantaneous load at time
t
=3
for a set of three real-time jobs.
9.1.2
TERMINOLOGY
When dealing with computational load, it is important to distinguish between
overload
and
overrun
.
Definition 9.1
A computing system is said to experience an
overload
when the compu-
tation time demanded by the task set in a certain interval of time exceeds the available
processing time in the same interval.
Definition 9.2
A task (or a job) is said to experience an
overrun
when exceeding its
expected utilization. An overrun may occur either because the next job is activated
before its expected arrival time (
activation overrun
), or because the job computation
time exceeds its expected value (
execution overrun
).
Search WWH ::
Custom Search