A fast predictive algorithm with idle reduction for heterogeneous system scheduling

Optical Grid that provides a way to integrate the geographical distributed resources for large- scale grid applications. This paper mainly focuses on the task scheduling issue in Optical Grid and proposes two task scheduling algorithms to mini-mize the total time or scheduling length required for a given task. A simulated optical grid has been addressed to demon-strate the effectiveness of the schedulin g approach. The performance tests and anal ysis of the two algorithms are investi-gated to show the effect of the routing policy on the schedulin g length. A more flexible policy is able to further improve scheduling length. Keywords: Optical Grid, Task Scheduling Algorithm, Communication Contention I INTRODUCTION Distributed computing, also known as computational Grid, is a technology that makes it possible to utilize the geographi-cal distributed computational resources to solve the intensive computation of large-scale data. Most Grid systems were built over the IP packet switched networks in the past. In the situation when some data-intensive Grid applications re-quire moving Gigabytes or even larger of data, the low speed of the network becomes the bottleneck of Grid system, which could not allow the remote resources to be used efficiently. Nowadays the optical network with Synchronous Digital Hierarchy (SDH), Wavelength-D ivision Multiplexing (WDM), and Automa tic Switched Optical Network (ASON) infrastructure can provide a much larger amount of bandwidth than the electronic network, such as IP packet switched network. Therefore, it is significant to use the optical network as the network infrastructure of the computational Grid. We refer to this infrastructure as Optical Grid, which employs optical network to interconnect grid computational re-sources (e.g. supercomputers, computer clusters, storage and display devices) with dynamically provisioned lightpaths [1]. It has been realized that task scheduling is a fundamental issue in achieving high performance in grid environments [6]. Task scheduling aims to optimize the overall performance of a system, e.g., minimizing the average job response time and maximizing the number of jobs exec uted in certain period of time. Optical grid task scheduling is quite different from classic grid task scheduling model. In Optical Grid system, we can treat the optical network as a kind of resource, the scheduling algorithm in Optical Grid system should schedule the tasks to appropriate resources as we ll as scheduling the communications between tasks to appropriate optical links. The objec-tive of task scheduling is to minimize the completion time (we called it as scheduling length or make span ) of an appli-cation. Usually, this problem is modeled and solved under several assumptions. Unless the assumptions are simplified enough, the problem has been proven to be NP-complete [8]. Therefore, many heuristic algorithms have been proposed. But these heuristics either are based on the assumption that the processors are fully connected, or just don’t take into ac-count the optical network’s characters, es pecially the characters of WDM optical network. While actually, in the Optical Grid system, the computational resources are more likely not fully connected, and the communication contention is an important factor that affects the completion time of a Grid application. Thus, a new heuris tic algorithm that can schedule tasks and also the communications are needed. In this paper, we mainly focus on the static scheduling problem, which means the character of the Grid applica tion, including the task execution times, task dependencies, task communications and synchronization are known a priori.

With the changes in the era of science and technology, the emergence of 5G technology makes a communication speed leap, and the delay is also getting shorter and shorter. At present, the computing power of computing equipment is also steadily improving. As one of its key technologies, mobile edge computing (MEC) is becoming more and more important. Here, we introduce MEC task scheduling and device collaboration algorithms based on deep learning. Aiming at the multi-task joint optimization scheduling strategy microtearing mode to minimize the total service cost of multiple tasks, we establish a theoretical optimization model of multi-task joint scheduling and use the model to design a multi-task cloud collaborative optimization scheduling algorithm based on Livepeer Token rules. We propose a deep-learning device collaboration algorithm based on alternating direction multiplier. The sub differential algorithm and the improved conjugate gradient method are designed to solve the sub problem respectively. This paper simulates the proposed deep-learning task scheduling algorithm and device collaboration algorithm. The simulation results show that, in the task scheduling algorithm, the proposed algorithm reduces the cost of completing the task, but in the total task time, this algorithm has a certain gap compared with other algorithms. In the equipment collaboration algorithm, with the increase of the n value, the algorithm converges in a certain iteration increment; in terms of algorithm convergence, the algorithm needs 142 iterations at most to achieve convergence, and the fastest speed only needs 17 iterations. In the experiment based on real data sets, the proposed algorithm is also better than the other two algorithms. From these results, we can see that the proposed algorithm has good convergence, stability, and scalability, and it has excellent performance in task scheduling and device collaboration of MEC.

SAAS parallel task scheduling based on cloud service flow load algorithm

Objectives: Current state-of-the-art task scheduling algorithms were mainly focused on deadline, load and energy factors in centralized cloud context. So, the proposed research objective focuses on dynamic and decentralized context. Methods/Statistical Analysis: Multi-objective task scheduling has become an important criterion for the dynamic and decentralized nature of cloud environment. Moreover, existing research works assumes that the resource load, energy and task execution time are known due its homogeneous nature. In order to improve the cloud consumer’s satisfaction, a novel Locality-Load-Prediction Aware Multi-objective Task Scheduling (LLPAMTS) algorithm is proposed to eventually distribute the tasks according to dynamic nature of cloud virtual machines. Findings: Proposed LLPAMTS algorithm will effectively schedule the tasks in an optimized manner by VM Scheduler component. This scheduling algorithm exploits the various monitoring parameters like locality, load and prediction parameters. It outperforms the existing deadline, load and energy aware scheduling algorithms in terms of task transfer time, task waiting time, task execution time, and task completion time. Applications/Improvements: The proposed LLPAMTS algorithm provides an average of 5 to 10% less task completion time compared to the existing deadline, load and energy aware scheduling algorithms. Keywords: Cloud Environment, Heterogeneous Cloud, Locality-Load-Prediction Aware Scheduling, Multi-Objective, Task Scheduling

In the past few years, with the rapid development of heterogeneous computing systems (HCS), the issue of energy consumption has attracted a great deal of attention. How to reduce energy consumption is currently a critical issue in designing HCS. In response to this challenge, many energy-aware scheduling algorithms have been developed primarily using the dynamic voltage-frequency scaling (DVFS) capability which has been incorporated into recent commodity processors. However, these techniques are unsatisfactory in minimizing both schedule length and energy consumption. Furthermore, most algorithms schedule tasks according to their average-case execution times and do not consider task execution times with probability distributions in the real-world. In realizing this, we study the problem of scheduling a bag-of-tasks (BoT) application, made of a collection of independent stochastic tasks with normal distributions of task execution times, on a heterogeneous platform with deadline and energy consumption budget constraints. We build execution time and energy consumption models for stochastic tasks on a single processor. We derive the expected value and variance of schedule length on HCS by Clark's equations. We formulate our stochastic task scheduling problem as a linear programming problem, in which we maximize the weighted probability of combined schedule length and energy consumption metric under deadline and energy consumption budget constraints. We propose a heuristic energy-aware stochastic task scheduling algorithm called ESTS to solve this problem. Our algorithm can achieve high scheduling performance for BoT applications with low time complexity $O(n(M+\log n))$ , where $n$ is the number of tasks and $M$ is the total number of processor frequencies. Our extensive simulations for performance evaluation based on randomly generated stochastic applications and real-world applications clearly demonstrate that our proposed heuristic algorithm can improve the weighted probability that both the deadline and the energy consumption budget constraints can be met, and has the capability of balancing between schedule length and energy consumption.

Grid computing promotes resource sharing, dynamic computational resource allocation, distributed data access from disjoint application domains and allowing various service providers to meet different demands efficiently. With an increase of user demands to the resources, planning the guaranteed Quality of Service (QoS) is a challenging task. The objective of this paper is to propose an agent based grid framework and QoS Time Based Scheduling (QTBS) algorithm for effective task scheduling in the grid computing paradigm. This algorithm is simulated in the GridSim toolkit and the results shows that the proposed work gives better results in makespan, resource utilization rate and load balancing level than algorithms such as QoS guided Weighted Mean Time Min (QWMTM), QoS guided MinMin, Max-Min and Min-Min heuristic algorithms.

In a heterogeneous system, processor and network failure are inevitable and can have adverse effect on the complex applications executing on the systems. To reduce the rate of these failures, matching and scheduling algorithms should take into account the objectives of minimizing schedule length (makespan) and reducing the probability of failure. Equitable distribution of workload over resources contributes in reducing the probability of failure. The Heterogeneous Earliest Finish Time (HEFT) algorithm has been proved as a performance effective task scheduling algorithm, addressing the objective of minimizing makespan. Reliable Dynamic Level Scheduling (RDLS) algorithm is a bi-objective scheduling algorithm that maximizes the reliability more effectively. Though the reliable version of HEFT algorithm (RHEFT) considers failure rate in scheduling decision, the improvement in reliability is less, compared to that of RDLS. To overcome this deficiency, we propose to incorporate the task–processor pair finding step of RDLS in HEFT algorithm, since it meets both the objectives of minimizing the makespan and maximizing the reliability. We define the load on a processor as the amount of time the processor is engaged in completing the scheduled subtasks. In this paper, a modification to the HEFT is proposed as a new algorithm called Improved Reliable HEFT (IRHEFT) for minimizing the schedule length, balancing the load and maximizing the reliability of schedule. The algorithm is compared for its performance with RDLS algorithm for randomly generated task graphs and a real application task graph.

In this paper, scheduling parallel tasks on multiprocessor computers with dynamically variable voltage and speed are addressed as combinatorial optimization problems. Two problems are defined, namely, minimizing schedule length with energy consumption constraint and minimizing energy consumption with schedule length constraint. The first problem has applications in general multiprocessor and multicore processor computing systems where energy consumption is an important concern and in mobile computers where energy conservation is a main concern. The second problem has applications in real-time multiprocessing systems and environments where timing constraint is a major requirement. Our scheduling problems are defined such that the energy-delay product is optimized by fixing one factor and minimizing the other. It is noticed that power-aware scheduling of parallel tasks has rarely been discussed before. Our investigation in this paper makes some initial attempt to energy-efficient scheduling of parallel tasks on multiprocessor computers with dynamic voltage and speed. Our scheduling problems contain three nontrivial subproblems, namely, system partitioning, task scheduling, and power supplying. Each subproblem should be solved efficiently, so that heuristic algorithms with overall good performance can be developed. The above decomposition of our optimization problems into three subproblems makes design and analysis of heuristic algorithms tractable. A unique feature of our work is to compare the performance of our algorithms with optimal solutions analytically and validate our results experimentally, not to compare the performance of heuristic algorithms among themselves only experimentally. The harmonic system partitioning and processor allocation scheme is used, which divides a multiprocessor computer into clusters of equal sizes and schedules tasks of similar sizes together to increase processor utilization. A three-level energy/time/power allocation scheme is adopted for a given schedule, such that the schedule length is minimized by consuming given amount of energy or the energy consumed is minimized without missing a given deadline. The performance of our heuristic algorithms is analyzed, and accurate performance bounds are derived. Simulation data which validate our analytical results are also presented. It is found that our analytical results provide very accurate estimation of the expected normalized schedule length and the expected normalized energy consumption and that our heuristic algorithms are able to produce solutions very close to optimum.

Parallel computing is an important area of research. It is used from scientific to commercial applications. Task scheduling is one of the important research areas of parallel computing. The major objective of any task scheduling algorithm is to minimize the execution time i.e. to reduce the scheduling length. Here, an application program is represented by a directed acyclic graph (DAG). In this paper, we have proposed an efficient CPTL scheduling algorithm which is based on two well-known attributes: Critical Path (CP), and Top-level (T-level). This algorithm is calculated the difference between CP and t-level, which is called CPT attribute. This attribute used as a priority of the tasks, and allocated these tasks on multiprocessor system. The proposed algorithm outperforms over the existing heuristic algorithms such as HLEFT, MCP, ETF, and DLS. CPTL algorithm gives minimum scheduling length. The comparison has been done based on following metrics: scheduling length, speedup, efficiency, load balancing, and normalized scheduling length.

Task scheduling is a well-known NP-complete problem. It is one of the research areas in parallel computing. Task scheduling consists of an application program and a set of processors for allocating the various tasks. An application program is represented using a directed acyclic graph(DAG) and a set of processors that are homogenous and fully connected. In this paper, we study different task duplication-based (TDB) scheduling algorithms and their priority attributes and time complexity. Also, we will discuss different types of performance metrics in task scheduling algorithms. These performance metrics help to analyze different task scheduling algorithms.

Cloud computing has spread fast because of its high performance distributed computing. It offers services and access to shared resources to internet users through service providers. Efficient performance of task scheduling in clouds is one of the most important research issues which needs to be focused on. Various task scheduling algorithms for cloud based on metaheuristic techniques have been examined and showed high performance in reasonable time such as scheduling algorithms based on Ant Colony Optimization (ACO), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO). In this paper, we propose a new task-scheduling algorithm based on Lion Optimization Algorithm (LOA), for cloud computing. LOA is a nature-inspired population-based algorithm for obtaining global optimization over a search space. It was proposed by Maziar Yazdani and Fariborz Jolai in 2015. It is a metaheuristic algorithm inspired by the special lifestyle of lions and their cooperative characteristics. The proposed task scheduling algorithm is compared with scheduling algorithms based on Genetic Algorithm and Particle Swarm Optimization. The results demonstrate the high performance of the proposed algorithm, when compared with the other algorithms.

Efficient scheduling algorithms are key for attaining high performance in heterogeneous computing systems. In this article, we propose a new list scheduling algorithm for assigning task graphs to fully connected heterogeneous processors with an aim to minimize the scheduling length. The proposed algorithm, called Improved Predict Priority Task Scheduling (IPPTS) algorithm has two phases: task prioritization phase, which gives priority to tasks, and processor selection phase, which selects a processor for a task. The IPPTS algorithm has a quadratic time complexity as the related algorithms for the same goal, that is <inline-formula><tex-math notation="LaTeX">$O(t^{2} \times p)$</tex-math></inline-formula> , for <inline-formula><tex-math notation="LaTeX">$t$</tex-math></inline-formula> tasks and <inline-formula><tex-math notation="LaTeX">$p$</tex-math></inline-formula> processors. Our algorithm reduces the scheduling length significantly by looking ahead in both task prioritization phase and processor selection phase. In this way, the algorithm is looking ahead to schedule a task and its heaviest successor task to the optimistic processor, i.e., the processor that minimizes their computation and communication costs. The experiments based on both randomly generated graphs and graphs of real-world applications show that the IPPTS algorithm significantly outperforms previous list scheduling algorithms in terms of makespan, speedup, makespan standard deviation, efficiency, and frequency of best results.

Failure-resilient DAG task scheduling in edge computing

Wireless Cloud computing delivers the data and computing resources through the internet, on a pay for usage basis. By using this, we can automatically update our software. We can use only the space required for the server, which reduces the carbon footprint. Task scheduling is the main problem in cloud computing which reduces the system performance. To improve system performance, there is need of an efficient task-scheduling algorithm. Existing task-scheduling algorithms focus on task-resource requirements, CPU memory, execution time and execution cost. However, these do not consider network bandwidth. In this paper, we introduce an efficient task-scheduling algorithm, which presents divisible task scheduling by considering network bandwidth. By this, we can allocate the workflow based on the availability of network bandwidth. Our proposed task-scheduling algorithm uses a nonlinear programming model for divisible task scheduling, which assigns the correct number of tasks to each virtual machine. Based on the allocation, we design an algorithm for divisible load scheduling by considering the network bandwidth.

Cloud computing is becoming very common in recent years and is growing rapidly due to its attractive benefits and features such as resource pooling, accessibility, availability, scalability, reliability, cost saving, security, flexibility, on-demand services, pay-per-use services, use from anywhere, quality of service, resilience, etc. With this rapid growth of cloud computing, there may exist too many users that require services or need to execute their tasks simultaneously by resources provided by service providers. To get these services with the best performance, and minimum cost, response time, makespan, effective use of resources, etc. an intelligent and efficient task scheduling technique is required and considered as one of the main and essential issues in the cloud computing environment. It is necessary for allocating tasks to the proper cloud resources and optimizing the overall system performance. To this end, researchers put huge efforts to develop several classes of scheduling algorithms to be suitable for the various computing environments and to satisfy the needs of the various types of individuals and organizations. This research article provides a classification of proposed scheduling strategies and developed algorithms in cloud computing environment along with the evaluation of their performance. A comparison of the performance of these algorithms with existing ones is also given. Additionally, the future research work in the reviewed articles (if available) is also pointed out. This research work includes a review of 88 task scheduling algorithms in cloud computing environment distributed over the seven scheduling classes suggested in this study. Each article deals with a novel scheduling technique and the performance improvement it introduces compared with previously existing task scheduling algorithms. Keywords: Cloud computing, Task scheduling, Load balancing, Makespan, Energy-aware, Turnaround time, Response time, Cost of task, QoS, Multi-objective. DOI: 10.7176/IKM/12-5-03 Publication date: September 30 th 2022