Hard-real-time scheduling of data-dependent tasks in embedded streaming applications

Abstract

Most of the hard-real-time scheduling theory for multiprocessor systems assumes independent periodic or sporadic tasks. Such a simple task model is not directly applicable to modern embedded streaming applications. This is because a modern streaming application is typically modeled as a directed graph where nodes represent actors (i.e. tasks) and edges represent data-dependencies. The actors in such graphs have data-dependency constraints and do not necessarily conform to the periodic or sporadic task models. Therefore, in this paper we investigate the applicability of hardreal-time scheduling theory for periodic tasks to streaming applications modeled as acyclic Cyclo-Static Dataflow (CSDF) graphs. In such graphs, the actors are data-dependent, however, we analytically prove that they (i.e. the actors) can be scheduled as implicitdeadline periodic tasks. As a result, a variety of hard-real-time scheduling algorithms for periodic tasks can be applied to schedule such applications with a certain guaranteed throughput. We compare the throughput resulting from such scheduling approach to the maximum achievable throughput of an application for a set of 19 real streaming applications. We find that in more than 80% of the cases, the throughput resulting from our approach is equal to the maximum achievable throughput.

Generalized Extraction of Real-Time Parameters for Homogeneous Synchronous Dataflow Graphs

“…In [10]- [12], the authors provide an analytical framework for computing timing parameters for actors of acyclic Cyclo-Static Dataflow (CSDF) applications with single-input streaming. The actors are considered as implicit-deadline and constrained-deadline periodic tasks in [10] and [11], [12], respectively.…”

“…The actors are considered as implicit-deadline and constrained-deadline periodic tasks in [10] and [11], [12], respectively. In contrast, the proposed approach is more general and can deal with any HSDF graph (CSDF can be converted to an HSDF), single/multiple input, and actors are modelled as arbitrary-deadline tasks.…”

“…The real-time dataflow applications are represented as cyclic HSDF with periodic sources. Earlier works have already considered applications represented as Directed Acyclic Graphs (DAGs) [10], [17], pipelines [18], task model without deadlines [15] or limited to implicit-deadlines [10], or application specific analysis [15], [19]. A main advantage of this proposal is that the extraction of the timing parameters is independent of the specific scheduler being used, or of other applications running in the system or the particular platform.…”

Resource optimization for CSDF-modeled streaming applications with latency constraints

“…In [4], a real-time scheduling framework for CSDF graphs is proposed. In this framework, every actor in a CSDF graph is characterized by a 4-tuple τ j = {S j , C j , D j , T j } in which S j is the start time, C j is the worst-case execution time (WCET), D j is the relative deadline, and T j is the period.…”

“…In addition, the latency of the CSDF graph can be reduced by selecting proper deadlines of actors. Moreover, [4] and [5] showed that a class of CSDFs scheduled as a periodic task set can achieve the maximum throughput and minimum latency. Scheduling a CSDF graph as a periodic task set makes it possible to benefit from a large amount of proven theories developed by the real-time systems community, which provide several efficient and fast approaches to compute the number of processors instead of performing complex design space exploration.…”