Energy efficient scheduling for parallel applications on mobile clusters

During the past decade, cluster computing and mobile communication technologies have been extensively deployed and widely applied because of their giant commercial value. The rapid technological advancement makes it feasible to integrate these two technologies and a revolutionary application called mobile cluster computing is arising on the horizon. Mobile cluster computing technology can further enhance the power of our laptops and mobile devices by running parallel applications. However, scheduling parallel applications on mobile clusters is technically challenging due to the significant communication latency and limited battery life of mobile devices. Therefore, shortening schedule length and conserving energy consumption have become two major concerns in designing efficient and energy-aware scheduling algorithms for mobile clusters. In this paper, we propose two novel scheduling strategies aimed at leveraging performance and power consumption for parallel applications running on mobile clusters. Our research focuses on scheduling precedence constrained parallel tasks and thus duplication heuristics are applied to schedule parallel tasks to minimize communication overheads. However, existing duplication algorithms are developed with consideration of schedule lengths, completely ignoring energy consumption of clusters. In this regard, we design two energy-aware duplication scheduling algorithms, called EADUS and TEBUS, to schedule precedence constrained parallel tasks with a complexity of O(n ²), where n is the number of tasks in a parallel task set. Unlike the existing duplication-based scheduling algorithms that replicate all the possible predecessors of each task, the proposed algorithms judiciously replicate predecessors of a task if the duplication can help in conserving energy. Our energy-aware scheduling strategies are conducive to balancing scheduling lengths and energy savings of a set of precedence constrained parallel tasks. We conducted extensive experiments using both synthetic benchmarks and real-world applications to compare our algorithms with two existing approaches. Experimental results based on simulated mobile clusters demonstrate the effectiveness and practicality of the proposed duplication-based scheduling strategies. For example, EADUS and TABUS can reduce energy consumption for the Gaussian Elimination application by averages of 16.08% and 8.1% with merely 5.7% and 2.2% increase in schedule length respectively.

Genetic Algorithm Based Scheduling of Meta-Tasks with Stochastic Execution Times in Heterogeneous Computing Systemst1

In this study, we address the meta-task scheduling problem in heterogeneous computing (HC) systems, which is to find a task assignment that minimizes the schedule length of a meta-task composed of several independent tasks with no data dependencies. The fact that the meta-task scheduling problem in HC systems is NP-hard has motivated the development of many heuristic scheduling algorithms. These heuristic algorithms, however, neglect the stochastic nature of task execution times in an attempt to minimize a deterministic objective function, which is the maximum of the expected values of machine loads. Contrary to existing heuristics, we account for this stochastic nature by modeling task execution times as random variables. We, then, formulate a stochastic scheduling problem where the objective is to minimize the expected value of the maximum of machine loads. We prove that this new objective is underestimated by the deterministic objective function and that an optimal task assignment obtained with respect to the deterministic objective function could be inefficient in a real computing platform. In order to solve the stochastic scheduling problem posed, we develop a genetic algorithm based scheduling heuristic. Our extensive simulation studies show that the proposed genetic algorithm can produce better task assignments as compared to existing heuristics. Specifically, we observe a performance improvement on the relative cost heuristic (M.-Y. Wu and W. Shu, A high-performance mapping algorithm for heterogeneous computing systems, in: Int. Parallel and Distributed Processing Symposium, San Francisco, CA, April 2001) by up to 61%.     

Dynamic data replacement and adaptive scheduling policies in spark

Improper data replacement and inappropriate selection of job scheduling policy are important reasons for the degradation of Spark system operation speed, which directly causes the performance degradation of Spark parallel computing. In this paper, we analyze the existing caching mechanism of Spark and find that there is still more room for optimization of the existing caching policy. For the task structure analysis, the key information of Spark tasks is taken out to obtain the data and memory usage during the task runtime, and based on this, an RDD weight calculation method is proposed, which integrates various factors affecting the RDD usage and establishes an RDD weight model. Based on this model, a minimum weight replacement algorithm based on RDD structure analyzing is proposed. The algorithm ensure that the relatively more valuable data in the data replacement process can be cached into memory. In addition, the default job scheduling algorithm of the Spark framework considers a single factor, which cannot form effective scheduling for jobs and causes a waste of cluster resources. In this paper, an adaptive job scheduling policy based on job classification is proposed to solve the above problem. The policy can classify job types and schedule resources more effectively for different types of jobs. The experimental results show that the proposed dynamic data replacement algorithm effectively improves Spark's memory utilization. The proposed job classification-based adaptive job scheduling algorithm effectively improves the system resource utilization and shortens the job completion time.

     

Aproximating static list schedules in dynamic multithreaded applications

List scheduling algorithms are known to be efficient when the application to be executed can be described statically as a Directed Acyclic Graph (DAG) of tasks. Regardless of knowing the entire DAG beforehand, obtaining an optimal schedule in a parallel machine is a NP-hard problem. Moreover, many programming tools propose the use of scheduling techniques based on list strategies. This paper presents an analysis of scheduling algorithms for multithread programs in a dynamic scenario where threads are created and destroyed during execution. We introduce an algorithm to convert DAGs, describing applications as tasks, into Directed Cyclic Graphs (DCGs) describing the same application designed in a multithread programming interface. Our algorithm covers case studies described in previous works, successfully mapping from the abstract level of graphs to the application environment. These mappings preserve the guarantees offered by the abstract model, providing efficient scheduling of dynamic programs that follow the intended multithread model. We conclude the paper presenting some performance results we obtained by list schedulers in dynamic multithreaded environments. We also compare these results with the best scheduling we could obtain with similar static task schedulers.     

An effective reliability-driven technique of allocating tasks on heterogeneous cluster systems

In large-scale heterogeneous cluster computing systems, processor and network failures are inevitable and can have an adverse effect on applications executing on such systems. One way of taking failures into account is to employ a reliable scheduling algorithm. However, most existing scheduling algorithms for precedence constrained tasks in heterogeneous systems only consider scheduling length, and not efficiently satisfy the reliability requirements of task. In recognition of this problem, we build an application reliability analysis model based on Weibull distribution, which can dynamically measure the reliability of task executing on heterogeneous cluster with arbitrary networks architectures. Then, we propose a reliability-driven earliest finish time with duplication scheduling algorithm (REFTD) which incorporates task reliability overhead into scheduling. Furthermore, to improve system reliability, it duplicates task as if task hazard rate is more than threshold \(\theta \) . The comparison study, based on both randomly generated graphs and the graphs of some real applications, shows that our scheduling algorithm can shorten schedule length and improve system reliability significantly.     

A task scheduling algorithm for arbitrarily-connected processors with awareness of link contention

In this paper, we present a new task scheduling algorithm, called Contention-Aware Scheduling (CAS) algorithm, with the objective of delivering good quality of schedules in low running-time by considering contention on links of arbitrarily-connected, heterogeneous processors. The CAS algorithm schedules tasks on processors and messages on links by considering the earliest finish time attribute with the virtual cut-through (VCT) or the store-and-forward (SAF) switching. There are three types of CAS algorithm presented in this paper, which differ in ordering the messages from immediate predecessor tasks. As part of the experimental study, the performance of the CAS algorithm is compared with two well-known APN (arbitrary processor network) scheduling algorithms. Experiments on the results of the synthetic benchmarks and the task graphs of the well-known problems clearly show that our CAS algorithm outperforms the related work with respect to performance (given in normalized schedule length) and cost (given in running time) to generate output schedules. Ali Fuat Alkaya received the B.Sc. degree in mathematics from Koc University, Istanbul, Turkey in 1998, and the M.Sc. degree in computer engineering from Marmara University, Istanbul, Turkey in 2002. He is currently a Ph.D. student in engineering management department at the same university. His research interests include task scheduling and analysis of algorithms. Haluk Rahmi Topcuoglu received the B.Sc. and M.Sc. degrees in computer engineering from Bogazici University, Istanbul, Turkey, in 1991 and 1993, respectively. He received the Ph.D. degree in computer science from Syracuse University in 1999. He has been on the faculty at Marmara University, Istanbul, Turkey since Fall 1999, where he is currently an Associate Professor in computer engineering department. His main research interests are task scheduling and mapping in parallel and distributed systems; parallel processing; evolutionary algorithms and their applicability for stationary and dynamic environments. He is a member of the ACM, the IEEE, and the IEEE Computer Society. e-mail: haluk@eng.marmara.edu.tr e-mail: falkaya@eng.marmara.edu.tr     

Energy and resource efficient workflow scheduling in a virtualized cloud environment

High energy consumption (EC) is one of the leading and interesting issue in the cloud environment. The optimization of EC is generally related to scheduling problem. Optimum scheduling strategy is used to select the resources or tasks in such a way that system performance is not violated while minimizing EC and maximizing resource utilization (RU). This paper presents a task scheduling model for scheduling the tasks on virtual machines (VMs). The objective of the proposed model is to minimize EC, maximize RU, and minimize workflow makespan while preserving the task’s deadline and dependency constraints. An energy and resource efficient workflow scheduling algorithm (ERES) is proposed to schedule the workflow tasks to the VMs and dynamically deploy/un-deploy the VMs based on the workflow task’s requirements. An energy model is presented to compute the EC of the servers. Double threshold policy is used to perceive the server’ status i.e. overloaded/underloaded or normal. To balance the workload on the overloaded/underloaded servers, live VM migration strategy is used. To check the effectiveness of the proposed algorithm, exhaustive simulation experiments are conducted. The proposed algorithm is compared with power efficient scheduling and VM consolidation (PESVMC) algorithm on the accounts of RU, energy efficiency and task makespan. Further, the results are also verified in the real cloud environment. The results demonstrate the effectiveness of the proposed ERES algorithm.

     

SEParAT: scheduling support environment for parallel application task graphs

Programs using parallel tasks can be represented by task graphs so that scheduling algorithms can be used to find an efficient execution order of the parallel tasks. This article proposes a flexible, component-based and extensible scheduling framework called SEParAT that supports the scheduling of a parallel program in multiple ways. The article describes the functionality and the software architecture of SEParAT. The flexible interfaces enable the cooperation with other programming tools, e.g., tools exploiting a specification of the parallel task structure of an application. The core component of SEParAT is an extensible scheduling algorithm library that provides an infrastructure to determine efficient schedules for task graphs. Homogeneous as well as heterogeneous platforms can be handled. The article also includes detailed experimental results comprising the evaluation of SEParAT as well as the evaluation of a variety of scheduling algorithms.     

A truthful online mechanism for virtual machine provisioning and allocation in clouds

We address the problem of online virtual machine (VM) provisioning and allocation with multiple types of resources. Formulating this problem in an auction-based setting, we propose an accurate mathematical model incorporating the ability to preempt and resume a given task for the sake of best overall use of resources. Our objective is to efficiently provide and allocate multiple VMs to maximize social welfare and encourage users to declare truthful requests. We first design an offline optimal mechanism based on the VCG mechanism; this mechanism has full knowledge of all users and offers ideal solutions. We also design an online greedy mechanism that considers only current knowledge while offering near-optimal solutions instead. Our proposed greedy mechanism consists of winner determination and payment algorithms. Furthermore, we show that the winner determination algorithm is monotonic and that the payment algorithm implements the critical payment. Both our allocation methods offer incentives to users providing true values for the sake of obtaining the best utility. We performed extensive experiments to investigate the performance of our proposed greedy mechanism compared to the optimal mechanism. Experimental results demonstrate that our proposed greedy mechanism obtains near-optimal solutions in a reasonable time.

     

Value of service based resource management for large-scale computing systems

Task scheduling for large-scale computing systems is a challenging problem. From the users perspective, the main concern is the performance of the submitted tasks, whereas, for the cloud service providers, reducing operation cost while providing the required service is critical. Therefore, it is important for task scheduling mechanisms to balance users’ performance requirements and energy efficiency because energy consumption is one of the major operational costs. We present a time dependent value of service (VoS) metric that will be maximized by the scheduling algorithm that take into consideration the arrival time of a task while evaluating the value functions for completing a task at a given time and the tasks energy consumption. We consider the variation in value for completing a task at different times such that the value of energy reduction can change significantly between peak and non-peak periods. To determine the value of a task completion, we use completion time and energy consumption with soft and hard thresholds. We define the VoS for a given workload to be the sum of the values for all tasks that are executed during a given period of time. Our system model is based on virtual machines, where each task will be assigned a resource configuration characterized by the number of the homogeneous cores and amount of memory. For the scheduling of each task submitted to our system, we use the estimated time to compute matrix and the estimated energy consumption matrix which are created using historical data. We design, evaluate, and compare our task scheduling methods to show that a significant improvement in energy consumption can be achieved when considering time-of-use dependent scheduling algorithms. The simulation results show that we improve the performance and the energy values up to 49% when compared to schedulers that do not consider the value functions. Similar to the simulation results, our experimental results from running our value based scheduling on an IBM blade server show up to 82% improvement in performance value, 110% improvement in energy value, and up to 77% improvement in VoS compared to schedulers that do not consider the value functions.     

Link contention-constrained scheduling and mapping of tasks and messages to a network of heterogeneous processors

In this paper, we consider the problem of scheduling and mapping precedence-constrained tasks to a network of heterogeneous processors. In such systems, processors are usually physically distributed, implying that the communication cost is considerably higher than in tightly coupled multiprocessors. Therefore, scheduling and mapping algorithms for such systems must schedule the tasks as well as the communication traffic by treating both the processors and communication links as equally important resources. We propose an algorithm that achieves these objectives and adapts its task scheduling and mapping decisions according to the given network topology. Just like tasks, messages are also scheduled and mapped to suitable links during the minimization of the finish times of tasks. Heterogeneity of processors is exploited by scheduling critical tasks to the fastest processors. Our experimental study has demonstrated that the proposed algorithm is efficient and robust, and yields consistent performance over a wide range of scheduling parameters. This revised version was published online in July 2006 with corrections to the Cover Date.     

Reconstruction of ancestral genomic sequences using likelihood.

A challenging task in computational biology is the reconstruction of genomic sequences of extinct ancestors, given the phylogenetic tree and the sequences at the leafs. This task is best solved by calculating the most likely estimate of the ancestral sequences, along with the most likely edge lengths. We deal with this problem and also the variant in which the phylogenetic tree in addition to the ancestral sequences need to be estimated. The latter problem is known to be NP-hard, while the computational complexity of the former is unknown. Currently, all algorithms for solving these problems are heuristics without performance guarantees. The biological importance of these problems calls for developing better algorithms with guarantees of finding either optimal or approximate solutions.We develop approximation, fix parameter tractable (FPT), and fast heuristic algorithms for two variants of the problem; when the phylogenetic tree is known and when it is unknown. The approximation algorithm guarantees a solution with a log-likelihood ratio of 2 relative to the optimal solution. The FPT has a running time which is polynomial in the length of the sequences and exponential in the number of taxa. This makes it useful for calculating the optimal solution for small trees. Moreover, we combine the approximation algorithm and the FPT into an algorithm with arbitrary good approximation guarantee (PTAS). We tested our algorithms on both synthetic and biological data. In particular, we used the FPT for computing the most likely ancestral mitochondrial genomes of hominidae (the great apes), thereby answering an interesting biological question. Moreover, we show how the approximation algorithms find good solutions for reconstructing the ancestral genomes for a set of lentiviruses (relatives of HIV). Supplementary material of this work is available at www.nada.kth.se/~isaac/publications/aml/aml.html.     

MIP formulations for an application of project scheduling in human resource management

In the literature, various discrete-time and continuous-time mixed-integer linear programming (MIP) formulations for project scheduling problems have been proposed. The performance of these formulations has been analyzed based on generic test instances. The objective of this study is to analyze the performance of discrete-time and continuous-time MIP formulations for a real-life application of project scheduling in human resource management. We consider the problem of scheduling assessment centers. In an assessment center, candidates for job positions perform different tasks while being observed and evaluated by assessors. Because these assessors are highly qualified and expensive personnel, the duration of the assessment center should be minimized. Complex rules for assigning assessors to candidates distinguish this problem from other scheduling problems discussed in the literature. We develop two discrete-time and three continuous-time MIP formulations, and we present problem-specific lower bounds. In a comparative study, we analyze the performance of the five MIP formulations on four real-life instances and a set of 240 instances derived from real-life data. The results indicate that good or optimal solutions are obtained for all instances within short computational time. In particular, one of the real-life instances is solved to optimality. Surprisingly, the continuous-time formulations outperform the discrete-time formulations in terms of solution quality.     

Delays without mistakes: response time and error distributions in dual-task

Background

When two tasks are presented within a short interval, a delay in the execution of the second task has been systematically observed. Psychological theorizing has argued that while sensory and motor operations can proceed in parallel, the coordination between these modules establishes a processing bottleneck. This model predicts that the timing but not the characteristics (duration, precision, variability…) of each processing stage are affected by interference. Thus, a critical test to this hypothesis is to explore whether the qualitiy of the decision is unaffected by a concurrent task.

Methodology/Principal Findings

In number comparison–as in most decision comparison tasks with a scalar measure of the evidence–the extent to which two stimuli can be discriminated is determined by their ratio, referred as the Weber fraction. We investigated performance in a rapid succession of two non-symbolic comparison tasks (number comparison and tone discrimination) in which error rates in both tasks could be manipulated parametrically from chance to almost perfect. We observed that dual-task interference has a massive effect on RT but does not affect the error rates, or the distribution of errors as a function of the evidence.

Conclusions/Significance

Our results imply that while the decision process itself is delayed during multiple task execution, its workings are unaffected by task interference, providing strong evidence in favor of a sequential model of task execution.     

An Effective Logical Cache for a Clustered LRC-Based DSM System

As local-area workstation networks are widely available, the idea of offering a software distributed shared memory (SDSM) system across interconnects of clusters is quite an attractive alternative for compute-intensive applications. However, the higher cost of sending a message over an inter-cluster link compared to an intra-cluster one can limit applications' performance on a multi-cluster SDSM system. In this paper, we present the extensions that we have added to the SDSM TreadMarks, which provides the lazy release consistency (LRC) memory model, in order to adapt it to a loosely-coupled cluster-based platform. We have implemented a logical per-cluster cache that exploits cluster locality. By accessing the cache of its cluster, a processor can share data previously requested by a second processor of its cluster, thereby, minimizing, the cost of inter-cluster communication.     

Energy aware DAG scheduling on heterogeneous systems

We address the problem of scheduling directed a-cyclic task graph (DAG) on a heterogeneous distributed processor system with the twin objectives of minimizing finish time and energy consumption. Previous scheduling heuristics have assigned DAGs to processors to minimize overall run-time of the application. But applications on embedded systems, such as high performance DSP in image processing, multimedia, and wireless security, need schedules which use low energy too.     

A two-phase heuristic for the energy-efficient scheduling of independent tasks on computational grids

The sensitivity analysis of a Cellular Genetic Algorithm (CGA) with local search is used to design a new and faster heuristic for the problem of mapping independent tasks to a distributed system (such as a computer cluster or grid) in order to minimize makespan (the time when the last task finishes). The proposed heuristic improves the previously known Min-Min heuristic. Moreover, the heuristic finds mappings of similar quality to the original CGA but in a significantly reduced runtime (1,000 faster). The proposed heuristic is evaluated across twelve different classes of scheduling instances. In addition, a proof of the energy-efficiency of the algorithm is provided. This convergence study suggests how additional energy reduction can be achieved by inserting low power computing nodes to the distributed computer system. Simulation results show that this approach reduces both energy consumption and makespan.     

Intelligent scheduling of FMSs with inductive learning capability using neural networks

With the growing uncertainty and complexity in the manufacturing environment, most scheduling problems have been proven to be NP-complete and this can degrade the performance of conventional operations research (OR) techniques. This article presents a system-attribute-oriented knowledge-based scheduling system (SAOSS) with inductive learning capability. With the rich heritage from artificial intelligence (AI), SAOSS takes a multialgorithm paradigm which makes it more intelligent, flexible, and suitable than others for tackling complicated, dynamic scheduling problems. SAOSS employs an efficient and effective inductive learning method, a continuous iterative dichotomister 3 (CID3) algorithm, to induce decision rules for scheduling by converting corresponding decision trees into hidden layers of a self-generated neural network. Connection weights between hidden units imply the scheduling heuristics, which are then formulated into scheduling rules. An FMS scheduling problem is also given for illustration. The scheduling results show that the system-attribute-oriented knowledge-based approach is capable of addressing dynamic scheduling problems.     

Block‐cyclic redistribution over heterogeneous networks

Clusters of workstations and networked parallel computing systems are emerging as promising computational platforms for HPC applications. The processors in such systems are typically interconnected by a collection of heterogeneous networks such as Ethernet, ATM, and FDDI, among others. In this paper, we develop techniques to perform block-cyclic redistribution over P processors interconnected by such a collection of heterogeneous networks. We represent the communication scheduling problem using a timing diagram formalism. Here, each interprocessor communication event is represented by a rectangle whose height denotes the time to perform this event over the heterogeneous network. The communication scheduling problem is then one of appropriately positioning the rectangles so as to minimize the completion time of all the communication events. For the important case where the block size changes by a factor of K, we develop a heuristic algorithm whose completion time is at most twice the optimal. The running time of the heuristic is O(PK ²). Our heuristic algorithm is adaptive to variations in network performance, and derives schedules at run-time, based on current information about the available network bandwidth. Our experimental results show that our schedules always have communication times that are very close to optimal. This revised version was published online in July 2006 with corrections to the Cover Date.     

Task placement for crowd recognition in edge-cloud based urban intelligent video systems

Recently, edge-cloud has attracted much attention by its promising prospect in terms of facilitating the benefits of edge and cloud together. It is promising for urban video systems that require efficient and effective processing for their intelligent monitoring drives on various ends, like sky drones and land cameras. For instance, to support crowd recognition for public safety, the tasks to crowd recognition need to be placed into all processing nodes in the video systems for processing effectively. This is a challenging problem to facilitate the edge-cloud orchestrated scenarios. However, the variability of tasks based on their complexities is not considered fully in existing strategies. In this regard, we model and analyse task placement for crowd recognition in edge-cloud intelligent video systems. Then, our strategies are proposed which are referred to Node-Graph based Task Placement (NGTP) and Cluster-Graph based Task Placement (CGTP). Specifically, with the help of data dependencies, NGTP utilises the greedy approach with node graphs in the centralised way for general scenarios. Comparatively, CGTP utilises data dependency and similarity for task placing in the decentralised way for emergency scenarios. The experiments demonstrate the superior and effectiveness performance in forming tasks cost and running time of our proposed approaches.