Task Duplication Research Articles

Scheduling DAG-based applications in multicluster environments with background workload using task duplication

Grids or multicluster computing environments are becoming increasingly popular to both scientific and commercial applications. Process scheduling remains a central issue to be effectively resolved ...

A DAG Scheduling Algorithm Based on Selected Duplication of Precedent Tasks

A DAG Scheduling Algorithm Based on Selected Duplication of Precedent Tasks

Optimizing latency and throughput of application workflows on clusters

Scheduling, in many application domains, involves optimization of multiple performance metrics. For example, application workflows with real-time constraints have strict throughput requirements and also desire a low latency or response time. In this paper, we present a novel algorithm for the scheduling of workflows that act on a stream of input data. Our algorithm focuses on the two performance metrics, latency and throughput, and minimizes the latency of workflows while satisfying strict throughput requirements. We also describe steps to use the above approach to solve the problem of meeting latency requirements while maximizing throughput. We leverage pipelined, task and data parallelism in a coordinated manner to meet these objectives and investigate the benefit of task duplication in alleviating communication overheads in the pipelined schedule for different workflow characteristics. The proposed algorithm is designed for a realistic bounded multi-port communication model, where each processor can simultaneously communicate with at most k distinct processors. Experimental evaluation using synthetic benchmarks as well as those derived from real applications shows that our algorithm consistently produces lower latency schedules that meet throughput requirements, even when previously proposed schemes fail.

Process Scheduling in Heterogeneous Multiprocessor Systems Using Task Duplication

Scheduling tasks in heterogeneous parallel and distributed computing environments continues to be a challenging problem. In this paper, the authors investigate the Heterogeneous Earliest Finish Time (HEFT) algorithm, along with alternative scheduling policies for task prioritising phases and the Critical Path on a Processor (CPOP) for scheduling tasks on a heterogeneous multiprocessor system. The authors show that by combining the HEFT algorithm selection policy with the task duplication strategy, it is possible to further reduce the schedule length produced by both HEFT and CPOP. The process scheduling algorithm presented in this paper compares favourably with other algorithms that use a similar strategy. The proposed algorithm has a time complexity of ?(¦V¦2(p + d)), whererepresents the number of tasks, p represents the number of processors and d the maximum in-degree of tasks.

A Clustering and Scheduling Algorithm Based on Task Duplication

This paper addresses an important and classic scheduling problem,the static scheduling of dependent tasks in homogeneous environment.It is NP hard even when the resources are unbounded,and finds many applications in the parallel and distributed computation area.Dependent tasks are usually denoted by Directed Acyclic Graph(DAG),and solving heuristics are commonly categorized to priority list based,cluster based and task duplication based schemes.Task-duplication-based(TDB)algorithms are of better performance than non-duplication ones.A new TDB clustering and scheduling algorithm,called the dynamic critical predecessor(DCP)algorithm,is proposed in this paper.DCP algorithm defines a new selective strategy for important ancestors to be duplicated.The primary aim is to get the shortest schedule length,and the next is to utilize as less resources as possible.Based on an improved definition of granularity,DCP algorithm achieves a better performance guarantee for arbitrary DAG than relative works reported in the literature.Experimental results on several benchmarks show that DCP algorithm is quite effective and it exceeds other classic TDB algorithms.Especially for the classic EZ benchmark,DCP algorithm gets an optimal solution with 8 makespan,which is better than the optimal result taken for before with 8.5 makespan.Complement graph of a DAG is defined,and a similar algorithm is developed to produce a 2-optimal schedule for tree graph if task duplication is not allowed for the tasks.

Genetic algorithms for task scheduling problem

The scheduling and mapping of the precedence-constrained task graph to processors is considered to be the most crucial NP-complete problem in parallel and distributed computing systems. Several genetic algorithms have been developed to solve this problem. A common feature in most of them has been the use of chromosomal representation for a schedule. However, these algorithms are monolithic, as they attempt to scan the entire solution space without considering how to reduce the complexity of the optimization process. In this paper, two genetic algorithms have been developed and implemented. Our developed algorithms are genetic algorithms with some heuristic principles that have been added to improve the performance. According to the first developed genetic algorithm, two fitness functions have been applied one after the other. The first fitness function is concerned with minimizing the total execution time (schedule length), and the second one is concerned with the load balance satisfaction. The second developed genetic algorithm is based on a task duplication technique to overcome the communication overhead. Our proposed algorithms have been implemented and evaluated using benchmarks. According to the evolved results, it has been found that our algorithms always outperform the traditional algorithms.

Directed Acyclic Task Graph Scheduling for Heterogeneous Computing Systems by Dynamic Critical Path Duplication Algorithm

This paper addresses the static scheduling of a directed acyclic task graph (DAG) on a heterogeneous, bounded set of distributed processors to minimize the makespan. We first derive the lower and upper bounds on the makespan of assigning a given directed acyclic task graph on heterogeneous processors by deferent scheduling strategies. Based on the analysis, we present a new heuristic, known as Heterogeneous Dynamic Critical Path Duplication (HDCPD), for scheduling DAG on a set of heterogeneous processors. HDCPD assigns the tasks on the dynamic critical path to the suitable processors which minimize the earliest finish time for them, combining insertion-based scheduling and task duplication techniques. The comparison study by simulation on Simgrid, based on randomly generated DAG, shows that HDCPD surpasses previous approaches in terms of both quality and cost of schedules, which are mainly presented with schedule length, frequency of best result, and scheduling time metrics.

A novel task scheduling algorithm based on dynamic critical path and effective duplication for pervasive computing environment

AbstractIn order to effectively utilize massive heterogeneous resources and provide transparent computing capability to upper applications, task scheduling as the key issue of pervasive computing system becomes significantly important. Previous proposed priority and duplication based task scheduling algorithms, which can be applied in pervasive computing environment, usually have following limitations: critical path cannot be calculated accurately while neglecting the effect of resource availability in scheduling; in duplication based resource allocation stage, duplications without restriction would lead to some negative effects on final schedule length (SL). For the purpose of solving these problems, a novel task scheduling algorithm based on dynamic critical path (DCP) and effective duplication, called DCPED, is presented in this paper. In DCPED, a more accurate DCP calculation method which takes resource availability into account is introduced. Meanwhile an effective task duplication strategy is proposed to eliminate ineffective duplications and make an optimized schedule result by using space compression technique and dynamic critical path length (DCPL) based evaluation technique respectively. Finally, simulation results show that DCPED can outperform previous algorithms significantly in NSL and speedup rate metrics. Especially, it is very effective for utilizing computing resources and scheduling the fine‐grain and large‐scale workflow applications in pervasive computing system. Copyright © 2008 John Wiley & Sons, Ltd.

An evaluation of the integration of theory and practice during practice education: Placement A

Background: The need to increase student placement capacity resulted in a new model for the delivery of Placement A by the Scottish Stakeholder Partnership (Transition Team Implementation Programme, 2007). A number of initiatives to increase capacity have also been implemented in London, including a pilot building on the Scottish experience where Placement A is delivered in two parts: a series of integrated ‘day placements’, followed by a 2‐week block placement at an approved placement setting. The delivery of the 4 week model for Placement A was evaluated to obtain baseline data and to ensure students’ ability to integrate campus‐based learning with placement learning was considered in the design of an alternative delivery model.Methods: A questionnaire was completed by 35 third year undergraduate students at King's College London and 14 second year undergraduate students at London Metropolitan University, after they had returned to university after a 4‐week Placement A. Using a five‐point Likert scale, students were asked to indicate the level of agreement with 29 statements, including nine relating to the tasks and integration of learning between campus and placement. Other questions related to student enjoyment, motivation, range of experiences and student perception of Placement A's impact on their preparation for Placement B.Results: The results showed that 38 of the 53 (72%) students felt that the 4‐week placement duplicated learning at university, but the same number also felt that the tasks undertaken helped them to apply what they had learned at university. Whilst the majority of students (43/53) agreed that the placement was enjoyable, 26 students responded they were bored at times during their placement. Four students indicated that after Placement A they were rethinking their decision to study dietetics.Discussion: Students’ responses reflected the role of the placement in the application of theory, and the duplication of tasks between practice and university is likely to be necessary to fulfil this role. The integration of theory and practice is an important educational outcome of practice placements, and the finding that some students’ are unable to relate placement tasks to campus‐based learning raises a concern that the theory‐practice gap could be better addressed. Improved integration may benefit the student experience and enjoyment of Placement A.Conclusions: The results from this evaluation will inform the pilot of the new delivery model for Placement A in London.Reference Transition Team Implementation Programme. (2007) Working In Partnership Towards Local Allocation of Dietetic Practice Placements – TTIP Progress Report, February 2007. Available at http://www.lmu.ac.uk/health/ttip/documents/progressreport.pdf (accessed on 18 January 2008).

Improving Integrated General Medical and Mental Health Services in Community-based Practices

The historical fragmentation of physical and mental health services has impeded efforts to improve quality and outcomes of care for persons with mental disorders. However, there is little information on effective strategies that might reduce fragmentation and improve integrated services within non-academic, community-based healthcare settings. Twenty-three practices from across the U.S. participated in a learning community meeting designed to identify barriers to integrated care and strategies for reducing such barriers. Barriers were initially identified based on a quantitative survey of organizational factors. Focus groups were used to elaborate on barriers to integrated care and to identify strategies for reducing barriers that are feasible in community-based settings. Participants identified key barriers, including administrative (e.g., lack of common medical records for mental health and general medical conditions), financial (e.g., lack of reimbursement codes to bill for mental health and general medical care in the same setting), and clinical (e.g., lack of an integrated care protocol). Top strategies recommended by participants included templates (i.e., for memoranda of understanding) to allow providers to work across practice settings, increased medical record security to enable a common medical record between mental health and general medical care, working with state Medicaid agencies to establish integrated care reimbursement codes, and guidance in establishing workflows between different providers (i.e., avoid duplication of tasks). Strategies to overcome barriers to integrated care may require cooperation across different organizational levels, including administrators, providers, and health care payers in order for integrated care to be established and sustained over time.

A performance study of multiprocessor task scheduling algorithms

Multiprocessor task scheduling is an important and computationally difficult problem. A large number of algorithms were proposed which represent various tradeoffs between the quality of the solution and the computational complexity and scalability of the algorithm. Previous comparison studies have frequently operated with simplifying assumptions, such as independent tasks, artificially generated problems or the assumption of zero communication delay. In this paper, we propose a comparison study with realistic assumptions. Our target problems are two well known problems of linear algebra: LU decomposition and Gauss–Jordan elimination. Both algorithms are naturally parallelizable but have heavy data dependencies. The communication delay will be explicitly considered in the comparisons. In our study, we consider nine scheduling algorithms which are frequently used to the best of our knowledge: min–min, chaining, A*, genetic algorithms, simulated annealing, tabu search, HLFET, ISH, and DSH with task duplication. Based on experimental results, we present a detailed analysis of the scalability, advantages and disadvantages of each algorithm.

2.3.3 HCI Aspects of SysML and Architectural Frameworks

AbstractThe Human Computer Interface (HCI) is one of the most important aspects of any system. It governs how people perceive the environment in which the system is deployed, and can either enable or hinder their ability to interact with that environment. Specifying the appropriate characteristics of the interface is therefore crucial to the correct implementation of the system. The goals of HCI are to develop or improve the safety, utility, effectiveness, efficiency, and usability of systems that include computers. The challenges for HCI are to keep abreast of technology, and to ensure that their designs offer good HCI, as well as harnessing the potential functionality of the new technology. As SysML becomes more prevalent for modeling systems, integrating HCI aspects into these models has the potential for improving HCI, eliminating duplication of tasks, and making systems more useable. This paper will look at SysML and DoDAF/MoDAF (MAF) and how they contribute towards defining the parameters in which the HCI will take place.

Practical Scheduling of Bag-of-Tasks Applications on Grids with Dynamic Resilience

Over the past decade, the grid has emerged as an attractive platform to tackle various large-scale problems, especially in science and engineering. One primary issue associated with the efficient and effective utilization of heterogeneous resources in a grid is scheduling. Grid scheduling involves a number of challenging issues, mainly due to the dynamic nature of the grid. There are only a handful of scheduling schemes for grid environments that realistically deal with this dynamic nature that have been proposed in the literature. In this paper, two novel scheduling algorithms, called the shared-input-data-based listing (SIL) algorithm and the multiple queues with duplication (MQD) algorithm for bag-of-tasks (BoT) applications in grid environments are proposed. The SIL algorithm targets scheduling data-intensive BoT (DBoT) applications, whereas the MQD algorithm deals with scheduling computationally intensive BoT (CBoT) applications. Their common and primary forte is that they make scheduling decisions without fully accurate performance prediction information. Another point to note is that both scheduling algorithms adopt task duplication as an attempt to reduce serious schedule increases. Our evaluation study employs a number of experiments with various simulation settings. The results show the practicability and competitiveness of our algorithms when compared to existing methods

Scheduling Task Parallel Applications for Rapid Turnaround on Enterprise Desktop Grids

Desktop Grids are popular platforms for high throughput applications, but due to their inherent resource volatility it is difficult to exploit them for applications that require rapid turnaround. Efficient desktop Grid execution of short-lived applications is an attractive proposition and we claim that it is achievable via intelligent resource selection. We propose three general techniques for resource selection: resource prioritization, resource exclusion, and task duplication. We use these techniques to instantiate several scheduling heuristics. We evaluate these heuristics through trace-driven simulations of four representative desktop Grid configurations. We find that ranking desktop resources according to their clock rates, without taking into account their availability history, is surprisingly effective in practice. Our main result is that a heuristic that uses the appropriate combination of resource prioritization, resource exclusion, and task replication can achieve performance within a factor of 1.7 of optimal in practice.

An Efficient Clustering-Based Task Scheduling Algorithm for Parallel Programs with Task Duplication *

This paper presents an efficient task scheduling algorithm for multiprocessor systems based on clustering with task duplication. This algorithm has a relatively small time complexity of O(|E| log |V|) for a task graph of |V| nodes and |E| edges. Results from an extensive simulation run on randomly generated graphs as well as several application graphs demonstrate a significant improvement in using the proposed technique over several well-known fast techniques. Schedule lengths produced by the proposed algorithm are in general comparable to the ones generated by techniques with a much larger computation time requirement. Actual processing time required by this technique is among the smallest in all advanced techniques, while the number of processors needed is also among the least in all techniques.

Clustering and scheduling method based on task duplication

A new heuristic approach that resembles the evolution of interpersonal relationships in human society is put forward for the problem of scheduling multitasks represented by a directed acyclic graph. The algorithm includes dynamic-group, detach-graph and front-sink components. The priority rules used are new. Relationship number, potentiality, weight and merge degree are defined for cluster's priority, and task potentiality for tasks' priority. Experiments show the algorithm could get good result in short time. The algorithm produces another optimal solution for the classic MJD benchmark. Its average performance is better than five latter-day representative algorithms, especially six benchmarks of the nines.

A branch-and-bound based solution approach for the mixed-model assembly line-balancing problem for minimizing stations and task duplication costs

A common assumption in the literature on mixed-model assembly line balancing is that a task that is common to multiple models must be assigned to a single station. In this paper, we relax this restriction, and allow a common task to be assigned to different stations for different models. We seek to minimize the sum of costs of the stations and the task duplication. We develop an optimal solution procedure based on a backtracking branch-and-bound algorithm and evaluate its performance via a large set of experiments. A branch-and-bound based heuristic is then developed for solving large-scale problems. The heuristic solutions are compared with a lower bound and experiments show that the heuristic provides much better solutions than those obtained by traditional approaches.

Qualifying Use Case Diagram Associations

Object-oriented software engineering methodology defines systems requirements in the form of flow-oriented use cases. This OOSE process involves developing a formal problem description - a problem statement $then developing use cases to capture the functional requirements of a proposed system. According to the OOSE approach, use cases should be used throughout the entire development process. Thus, they play a major role in a project's analysis, design, testing, and implementation stages. Use cases have five key stakeholders namely customers, systems and business analysts, systems architects and developers, software testing and quality assurance engineers, and project managers. Conceptual use case diagrams that draw on critical analysis can enhance a project manager's ability to allocate medium-sized or large-scale development projects in a way that limits task duplication and functional overlaps

Scheduling directed a-cyclic task graphs on a bounded set of heterogeneous processors using task duplication

In a distributed computing environment, the schedule by which tasks are assigned to processors is critical to minimizing the overall run-time of the application. However, the problem of discovering the schedule that gives the minimum finish time is NP-Complete. This paper addresses static scheduling of a directed a-cyclic task graph (DAG) on a heterogeneous, bounded set of distributed processors to minimize the makespan. By combining several innovative techniques, including insertion-based scheduling and multiple task duplication, we present a new heuristic, known as Heterogeneous N-predecessor Decisive Path (HNDP), for scheduling directed a-cyclic weighted task graphs (DAGs) on a set of heterogeneous processors. We compare the performance of HNDP, under a range of varying input conditions, with two of the best existing heterogeneous heuristics namely HEFT and STDS. The results presented in this paper show that HNDP outperforms the two heuristics in terms of finish time and the number of processors employed over a wide range of parameters. The complexity of HNPD is O ( v 2 . p ) vs. O ( v 2 . p ) of HEFT and O ( v 2 ) of STDS where v is the number of nodes in the DAG.

An Optimal Scheduling Algorithm for Fork-Join Task Graphs

The Fork-Join structure is one of the basic modeling structures for parallel processing. Although some algorithms are able to find an optimal schedule under certain conditions, they ignore to economize processors and minimize the total completion time. This paper presents a Task Duplication based Balance Scheduling(TDBS) algorithm which can generate an optimal schedule for fork-join task graph with a complexity of O(vq+vlogv), where v and q are the number of tasks and processors respectively. By considering workload and idle time slots of the used processors, TDBS algorithm tries to assign tasks to scheduled processors and maximize their utilization. Simulation results show that TDBS algorithm has better speedup and efficiency than other compared algorithms. Therefore, TDBS algorithm is a viable option for practical high performance applications.