On Best-Effort Real-Time Assurances for Recovering from Distributable Thread Failures in Distributed Real-Time Systems

Ravindran, Binoy; Curley, Edward; Anderson, Jonathan; Jensen, E. Douglas

doi:10.1109/isorc.2007.45

Cited by 8 publications

(8 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We show that ACUA satisfies thread time constraints in the presence of crash failures and message losses, is early-deciding (i.e., its decision time is proportional to the actual number of crashes), and has a message and time complexity that compares favorably with other algorithms in its class. Furthermore, we show that ACUA has a better best-effort property -i.e., the affinity for feasibly completing as many high importance threads as possible, irrespective of thread urgency -than past thread scheduling algorithms [10,11]. We also prove the exception handling properties of ACUA.…”

Section: Introductionmentioning

confidence: 66%

“…In this paper, we consider the problem of scheduling threads in the presence of the previously mentioned uncertainties, focusing particularly on (arbitrary) node failures and message losses. Past efforts on thread scheduling (e.g., [9][10][11]) can be broadly categorized into two classes: independent node scheduling and collaborative scheduling. In the independent scheduling approach (e.g., [9,11]), threads are scheduled at nodes using propagated thread scheduling parameters and without any interaction with other nodes.…”

Section: Introductionmentioning

confidence: 99%

“…Past efforts on thread scheduling (e.g., [9][10][11]) can be broadly categorized into two classes: independent node scheduling and collaborative scheduling. In the independent scheduling approach (e.g., [9,11]), threads are scheduled at nodes using propagated thread scheduling parameters and without any interaction with other nodes. Fault-management is separately addressed by thread integrity protocols that run concurrent to thread execution.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Scheduling distributable real-time threads in the presence of crash failures and message losses

Fahmy

Ravindran

Jensen

2008

Proceedings of the 2008 ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

We consider the problem of scheduling distributable realtime threads under run-time uncertainties including those on thread execution times, thread arrivals, node failures, and message losses. We present a distributed scheduling algorithm called ACUA that is designed under a partially synchronous model, allowing for probabilistically-described message delays. We show that ACUA satisfies thread time constraints in the presence of crash failures and message losses, is early-deciding, and has an efficient message and time complexity. The algorithm has also better "best-effort" real-time property than past thread scheduling algorithms.

show abstract

Section: Introductionmentioning

confidence: 66%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Scheduling distributable real-time threads in the presence of crash failures and message losses

Fahmy

Ravindran

Jensen

2008

Proceedings of the 2008 ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Furthermore, unlike previous authors, we do not assume the expression of real-time constraints for the abort handler [39], [95]. This is done because each real-time segment may have many different abort handlers, and it is assumed that the cost of expressing time constraints for each of these to the underlying operating system is inordinately large, compared to the execution cost of the abort handler.…”

Section: Abort Modelmentioning

confidence: 99%

An experimental evaluation of the scalability of real-time scheduling algorithms on large-scale multicore platforms

Dellinger

Lindsay

Ravindran

2012

ACM J. Exp. Algorithmics

Self Cite

View full text Add to dashboard Cite

This thesis studies the problem of experimentally evaluating the scaling behaviors of existing multicore real-time task scheduling algorithms on large-scale multicore platforms. As chip manufacturers rapidly increase the core count of processors, it becomes imperative that multicore real-time scheduling algorithms keep pace. Thus, it must be determined if existing algorithms can scale to these new high core-count platforms. Significant research exists on the theoretical performance of multicore real-time scheduling algorithms, but the vast majority of this research ignores the effects of scalability. It has been demonstrated that multicore real-time scheduling algorithms are feasible for small core-count systems (e.g. 8-core or less), but thus far the majority of the algorithmic research has never been tested on high core-count systems (e.g. 48-core or more).We present an experimental analysis of the scalability of 16 multicore real-time scheduling algorithms. These algorithms include global, clustered, and partitioned algorithms. We cover a broad range of algorithms, including deadline-based and utility accrual scheduling algorithms. These algorithms are compared under metrics including schedulability, tardiness, deadline satisfaction ratio, and utility accrual ratio. We consider multicore platforms ranging from 8 to 48 cores. The algorithms are implemented in a real-time Linux kernel we create called ChronOS. ChronOS is based on the Linux kernel's PREEMPT RT patch, which provides the underlying operating system kernel with real-time capabilities such as full kernel preemptibility and priority inheritance for kernel locking primitives. ChronOS extends these capabilities with a flexible, scalable real-time scheduling framework.Our study shows that it is possible to implement global fixed and dynamic priority and simple global utility accrual real-time scheduling algorithms which will scale to large-scale multicore platforms. Interestingly, and in contrast to the conclusion of prior research, our results reveal that some global scheduling algorithms (e.g. G-NP-EDF) is actually scalable on large core counts (e.g. 48). In our implementation, scalability is restricted by lock contention over the global schedule and the cost of inter-processor communication, rather than the global task queue implementation. We also demonstrate that certain classes of utility accrual algorithms such as the GUA class are inherently not scalable. We show that algorithms implemented with scalability as a first-order implementation goal are able to provide real-time guarantees on our 48-core platform.

show abstract

“…Prior research on providing quality of service (QoS) using DOC middleware has addressed the timeliness [20] and availability [16,19] requirements of DRE systems. Moreover, several standards have defined interfaces and provide services and strategies to enhance the timeliness and fault-tolerance capabilities of DRE systems.…”

mentioning

confidence: 99%

Towards Middleware for Fault-Tolerance in Distributed Real-Time and Embedded Systems

Balasubramanian

Gokhale

Schmidt

et al. 2008

Distributed Applications and Interoperable Systems

View full text Add to dashboard Cite

Abstract. Distributed real-time and embedded (DRE) systems often require support for multiple simultaneous quality of service (QoS) properties, such as real-timeliness and fault tolerance, that operate within resource constrained environments. These resource constraints motivate the need for a lightweight middleware infrastructure, while the need for simultaneous QoS properties require the middleware to provide fault tolerance capabilities that respect time-critical needs of DRE systems. Conventional middleware solutions, such as Fault-tolerant CORBA (FT-CORBA) and Continuous Availability API for J2EE, have limited utility for DRE systems because they are heavyweight (e.g., the complexity of their feature-rich fault tolerance capabilities consumes excessive runtime resources), yet incomplete (e.g., they lack mechanisms that enable fault tolerance while maintaining real-time predictability). This paper provides three contributions to the development and standardization of lightweight real-time and fault-tolerant middleware for DRE systems. First, we discuss the challenges in realizing real-time faulttolerant solutions for DRE systems using contemporary middleware. Second, we describe recent progress towards standardizing a CORBA lightweight fault-tolerance specification for DRE systems. Third, we present the architecture of FLARe, which is a prototype based on the OMG real-time fault-tolerant CORBA middleware standardization efforts that is lightweight (e.g., leverages only those server-and client-side mechanisms required for real-time systems) and predictable (e.g., provides fault-tolerant mechanisms that respect time-critical performance needs of DRE systems).

show abstract

On Best-Effort Real-Time Assurances for Recovering from Distributable Thread Failures in Distributed Real-Time Systems

Cited by 8 publications

References 7 publications

Scheduling distributable real-time threads in the presence of crash failures and message losses

Scheduling distributable real-time threads in the presence of crash failures and message losses

An experimental evaluation of the scalability of real-time scheduling algorithms on large-scale multicore platforms

Towards Middleware for Fault-Tolerance in Distributed Real-Time and Embedded Systems

Contact Info

Product

Resources

About