Hardware Reference
In-Depth Information
J
new
t
J
1
t
J
2
t
J
3
t
J
4
t
t
0
Figure 2.15
Example of domino effect.
At time
t
0
, if job
J
new
were accepted, all the other jobs (previously schedulable)
would miss their deadlines. In planned-based algorithms, this situation is detected at
time
t
0
, when the guarantee is performed, and causes job
J
new
to be rejected.
In summary, the guarantee test ensures that, once a task is accepted, it will complete
within its deadline and, moreover, its execution will not jeopardize the feasibility of
the tasks that have been previously guaranteed.
BEST-EFFORT ALGORITHMS
In certain real-time applications, computational activities have soft timing constraints
that should be met whenever possible to satisfy system requirements. In these systems,
missing soft deadlines do not cause catastrophic consequences, but only a performance
degradation.
For example, in typical multimedia applications, the objective of the computing sys-
tem is to handle different types of information (such as text, graphics, images, and
sound) in order to achieve a certain quality of service for the users. In this case, the
timing constraints associated with the computational activities depend on the qual-
ity of service requested by the users; hence, missing a deadline may only affect the
performance of the system.
To efficiently support soft real-time applications that do not have hard timing require-
ments, a
best-effort
approach may be adopted for scheduling.
Search WWH ::
Custom Search