Top PDF A Scheduling Approach with Processor and Network Heterogeneity for Grid Environment

A Scheduling Approach with Processor and Network Heterogeneity for Grid Environment

A Scheduling Approach with Processor and Network Heterogeneity for Grid Environment

Abstract—   Processor heterogeneity is an important issue in grid environment. In this paper, a list based task scheduling algorithm, called “critical path scheduling with t-level” (CPST) for grid computing system is proposed. There are no. of scheduling algorithms such as HEFT [1] use mean execution time based b-level for task priority and SHCP [2] use task priority based on simple critical path. In CPST, a critical path based task sequence is generated with t-level value of tasks, where variance based computation and communication cost is used. The experimental results show that CPST algorithm performs better than HEFT, SHCP and HHS algorithm in grid environment for task graphs.
Mostrar mais

7 Ler mais

TAAG: An Efficient Task Allocation Algorithm for Grid Environment

TAAG: An Efficient Task Allocation Algorithm for Grid Environment

Abstract— Grid computing is a form of distributed computing where the resources of various computers are shared to solve a particular problem. Due to heterogeneity of resources, scheduling a task is significantly complex in nature. Scheduling strategy plays a vital role in the grid environment to schedule the user tasks and dispatch them to the appropriate grid resources. A good task scheduling method is the one which reduces the total time taken for execution of a given task in the grid. In this paper, we propose a new scheduling algorithm called TAAG (An Efficient Task Allocation Algorithm for Grid) for efficient allocation of tasks on resources in the Grid environment. The proposed algorithm tries to use the advantages of the recent popular task scheduling algorithm known as “Improved Min-Min Task Scheduling Algorithm” [2] and covers its disadvantages by using a new resource allocation model. In this paper, the loads of the task are divided and are allocated into resources with respect to the computing capacity of the resource. Experimental results show that the proposed algorithm obtain better solution in terms of completion time, cost, makespan and load balancing compared to the existing algorithms.
Mostrar mais

9 Ler mais

Grid and Data Analyzing and Security

Grid and Data Analyzing and Security

s it stands, there are lots of researches ve been done regarding using calculation systems based on computer networks. According to their importance and their effects on different aspects of calculation systems, in this regards several researches have been done. In these studies different dimensions of them were studied such as: Evaluation of Job-Scheduling Strategies for Grid Computing (Hamscher et al, 2000), UNICORE: A Grid Computing Environment (Erwin, 2001), Intrinsic vulnerability assessment of the aquifer in the Riana spring catchment by the method SINTACS (Janza and Prestor, 2002), Economic models for resource management and scheduling in Grid computing (Buyyal, 2003), Grid Computing: A Brief Technology Analysis (Smith, 2004), Trusted Grid Computing with Security Binding and Trust Integration (Song, Hwang and Kwok, 2005), Scheduling Algorithms for Grid Computing: State of the Art and Open Problems (Dong and Akl,2006), Introducing Virtual Private Overlay Network services in large scale Grid infrastructures (Palmieri, 2007), Implementation of Computational Grid
Mostrar mais

12 Ler mais

Heuristic Based Task Scheduling In Grid

Heuristic Based Task Scheduling In Grid

Abstract — Grid computing is concerned with coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations. Efficient scheduling of complex applications in a grid environment reveals several challenges due to its high heterogeneity, dynamic behavior and space shared utilization. Objectives of scheduling algorithms are increase in system throughput, efficiency and reduction in task completion time. The main focus of this paper is to highlight the merits of resource and task selection technique based on certain heuristics.
Mostrar mais

7 Ler mais

AN EFFICIENT FAULT TOLERANT SCHEDULING APPROACH FOR COMPUTATIONAL GRID

AN EFFICIENT FAULT TOLERANT SCHEDULING APPROACH FOR COMPUTATIONAL GRID

Grid computing serves as an important technology to facilitate distributed computation computational grids solve large scale scientific problems using heterogeneous geographically distributed resources. Problems like dispatching and scheduling of tasks are considered as major issues in computational grid environment. The Grid Scheduler must select proper resources for executing the tasks with less response time. There are various reasons such as network failure, overloaded resource conditions, or non-availability of required software components for execution failure. Thus, fault-tolerant systems should be able to identify and handle failures and support reliable execution in the presence of failures. Hence the integration of fault tolerance measures and communication time with scheduling gains much importance. In this study, a new fault tolerance based scheduling approach Fault Tolerant Min-Min (FTMM) for scheduling statically available meta tasks is proposed wherein failure rate and the fitness value are calculated. The performance of the fault tolerant scheduling policy is compared with min-min scheduling policy using GridSim and the results shows that the proposed policy performs better with less makespan in the presence of failures. The number of tasks successfully completed is also more when compared to the non-fault tolerant min-min scheduling policy. Thus the proposed FTMM algorithm not only achieves better hit rate but also improved makespan.
Mostrar mais

6 Ler mais

A SLA-Aware Scheduling Architecture in Grid System Using Learning Techniques

A SLA-Aware Scheduling Architecture in Grid System Using Learning Techniques

Computing intensive jobs and data intensive jobs handled differently, and local and global updates are used to obtain the most recent status of grid resources to schedule jobs more effectively in real time. System ModelSyed Nasir Mehmood Shah et al. [10] propose an algorithm for CPU scheduling of a modern multiprogramming operating system, design and development of new CPU scheduling algorithms (the Hybrid Scheduling Algorithm and the Dual Queue Scheduling Algorithm) with a view to minimize overall task schedule. The following paper extends this prioritized round robin heuristic from a single system multiprogramming environment, onto a multi- processor distributed architecture. As each scheduling strategy optimizes some of performance parameters such as making span, resource utilization, response time, workload balancing, service time, reliability, fairness deviation and throughput, we propose a SLA- aware scheduling model to achieve four important parameters such as resource utilization, response time, workload balancing and throughput.
Mostrar mais

10 Ler mais

Dual Objective Security Driven Scheduling Model for Computational Grid using GA

Dual Objective Security Driven Scheduling Model for Computational Grid using GA

A computational grid is a collection of geographically dispersed heterogeneous computing resources, giving the image of a single large virtual computing system to users [1] [2] [3]. Scheduling on such platform is an important and complex task more so being grid a heterogeneous system. The main challenge for job scheduling in grid is its highly dynamic environment, in which computing resources have their own access policies, security, availability etc. At the same time, resources are of greater heterogeneity in terms of their architectural design that includes desktop PCs to supercomputers. Thus, grid which is privately owned, heterogeneous, non dedicated network of computers utilizes the idle time of thousands of computing devices to harness the high performance computing power. From the users’ point of view, the objective of the grid computing is to simplify distributed heterogeneous computing in the manner the World Wide Web has simplified information sharing over the Internet [4]. The key factors in making the grid computing feasible are: the evolution of standards such as TCP/IP and Ethernet; the ever-increasing bandwidth of networks; the increasing availability of idle CPU cycles on
Mostrar mais

9 Ler mais

Representing Job Scheduling for Volunteer Grid Environment using Online Container Stowage

Representing Job Scheduling for Volunteer Grid Environment using Online Container Stowage

Abstract —Volunteer grid computing comprises of volunteer resources which are unpredictable in nature and as such the scheduling of jobs among these resources could be very uncertain. It is also difficult to ensure the successful completion of submit- ted jobs on volunteer resources as these resources may opt to withdraw from the grid system anytime or there might be a re- source failure, which requires job reassignments. However, a careful consideration of future jobs can make scheduling of jobs more reliable on volunteer resources. There are two possibilities; either to forecast the future jobs or to forecast the resource availability by studying the history events. In this paper an attempt has been made to utilize the future job forecasting in improving the job scheduling experience with volunteer grid re- sources. A scheduling approach is proposed that uses container stowage to allocate the volunteer grid resources based on the jobs submitted. The proposed scheduling approach optimizes the number of resources actively used. The approach presents online container stowage adaptability for scheduling jobs using volun- teer grid resources. The performance has been evaluated by making comparison to other scheduling algorithms adopted in volunteer grid. The simulation results have shown that the pro- posed approach performs better in terms of average turnaround and waiting time in comparison with existing scheduling algo- rithms. The job load forecast also reduced the number of job reassignments.
Mostrar mais

9 Ler mais

Bandwidth-Adaptive Scheduling for Quality of Service Enhancement of Real-Time Multimedia Applications in Network Processor Based Router

Bandwidth-Adaptive Scheduling for Quality of Service Enhancement of Real-Time Multimedia Applications in Network Processor Based Router

Abstract: Problem statement: An explosive growth of multimedia applications in internet has stressed the performance of routers. Hence managing Quality of Service (QoS) enhancement of real- time multimedia applications over IP is a significant and demanding challenge. Approach: To address this issue, Bandwidth Adaptive Stratified Round Robin (BASRR) packet scheduling algorithm has been proposed in this paper for enhancing quality of service of real-time multimedia applications. Embedded Network Processors (NP) have recently emerged with flexibility and speed to reduce the stress of the router by effectively processing the packets. The main objective of this study was to implement the proposed packet scheduling algorithm in a Network Processor (NP) based router for enhancing quality of service of real-time multimedia applications Results: The effectiveness of the BASRR algorithm has been verified by simulations using Intel’s IXP 2400 network processor. The results show that BASRR achieves about 71.25% reduction in jitter compared to SRR when the traffic has uniform distribution of real-time flows and non real-time flows. The reduction in average queuing delay is about 30% compared to SRR for all the types of traffic. Conclusion: The QoS for multimedia applications has been achieved by the proposed non-preemptive Bandwidth Adaptive Stratified Round Robin (BASRR) scheduling algorithm and outperforms the three well-known scheduling algorithms including DRR, WDRR and SRR. The results showed that BASRR is efficient with per packet complexity of O(1) and provides better fairness and reduced delay.
Mostrar mais

7 Ler mais

Min-min approach for scheduling in grid environment

Min-min approach for scheduling in grid environment

Xiaoyong Tang et al.[10] proposed a stochastic scheduling algorithm for precedence constrained tasks on Grid. He addressed the problems in scheduling a precedence constrained tasks of parallel application with random tasks processing time and edges communication time on Grid computing systems so as to minimize the makes pan in stochastic environment. This is a difficult problem and few efforts have been reported on its solution in the literature. The problem is first formulated in a form of stochastic scheduling model on Grid systems. Then, a stochastic heterogeneous earliest finish time (SHEFT) scheduling algorithm is developed that incorporates the expected value and variance of stochastic processing time into scheduling. The HEFT algorithm [10] selects the task with the highest upward rank (an upward rank is defined as the maximum distance from the current node to the exiting node, including the computational cost and communication cost) at each step. The selected task is then assigned to the processor which minimizes its earliest finish time with an insertion-based approach which considers the possible insertion of a task in an earliest idle time slot between two already-scheduled tasks on the same resource.
Mostrar mais

9 Ler mais

An Optimal Scheduling Algorithm for Real Time Applications in Grid System

An Optimal Scheduling Algorithm for Real Time Applications in Grid System

The objective of the proposed work is to use an optimal scheduling algorithm for real-time application. A grid is considered to be an infrastructure that bonds and unifies globally remote and diverse resources in order to provide computing support for a wide range of applications. Real time applications in an industrialized technological infrastructure such as telecommunication systems, factories, defense systems, aircraft and space stations pose relatively rigid requirements on their performance. Aircraft scheduling represents the best example of real-time applications. The main focus of this work is to check the time taken for turn-around activities which comprises of taxi in, load/unload baggage, deboarding, water fueling, cleaning, catering, boarding, de-icing, take off processes, thus relating in the lowest flight delays and shortest waiting time. The optimal scheduling algorithm is used for aircraft take-offs. The penalties are associated with proper scheduling but delayed turn around activities, improper scheduling and early/late takeoffs.
Mostrar mais

6 Ler mais

On the research of flow around obstacle using the viscous Cartesian grid technique

On the research of flow around obstacle using the viscous Cartesian grid technique

A new 2-D viscous Cartesian grid is proposed in current research. It is a combi- nation of the existent body-fitted grid and Cartesian grid technology. On the in- terface of the two different type of grid, a fined triangular mesh is used to connect the two grids. Tests with flow around the cylinder and aerofoil NACA0012 show that the proposed scheme is easy for implement with high accuracy.

4 Ler mais

SYNCHRONIZATION OF NATIONAL GRID NETWORK WITH THE ELECTRICITY SHIPS NETWORK IN THE "SHORE TO SHIP" SYSTEM

SYNCHRONIZATION OF NATIONAL GRID NETWORK WITH THE ELECTRICITY SHIPS NETWORK IN THE "SHORE TO SHIP" SYSTEM

In   the  vessel  power  systems  the  synchroniza on  auto‐ ma on  system has to keep up with the changes in the ves‐ sel  network. As men oned before, vessel network is a ‘so ’  network  that is characterized by frequent voltage and fre‐ quency  changes (connec ng and disconnec ng of receivers  of  high power comparable to the power of a power gener‐ a ng  system). Synchroniza on among autonomous auxilia‐ ry  engines in vessel systems is performed automa cally and  is   realized  by  synchronisers.  More  or  less  technologically  advanced   synchronizers  used  in  shipbuilding  guarantee  automa c   synchroniza on  in  minimal  me.  Synchronizer  T4500   by  SELCO  (a  company  well‐known  in  shipbuilding)  can  serve as an example of such synchroniser [6]. 
Mostrar mais

5 Ler mais

Constructing Student Personal Course Scheduling Problem with Spreading Activation on a Course Network

Constructing Student Personal Course Scheduling Problem with Spreading Activation on a Course Network

Nozawa et. al. had been proposed a syllabus analysis system[2]. In designing of an original curriculum by a higher education institution, or in external evaluation of an institution’s curriculum, comprehending the curricu- lum contents of many institutions in the same field is necessary. However, this is very hard to do even for edu- cation experts in that field. This system uses clustering view from each class syllabus description. The main aim of this system is not for students, but for teachers to make specialized curriculum. This system treats syllabus data which constitute some curriculum and are expressed in a common format, calculates the similarity between the syllabi based on the occurrence frequency of technical terms, clusters the syllabus, and thus helps to find dis- tinguishing features of the curriculum by visualizing and comparing the assignments of the syllabus to the clusters along various classification axes. Visualizing of syllabus data is useful to make student’s study strategy, but it is not always the case the students find suitable course plan because of their lack of desire. We have tried to make a model of the student’s interests.
Mostrar mais

6 Ler mais

Load Balancing in Partner-Based Scheduling Algorithm for Grid Workflow

Load Balancing in Partner-Based Scheduling Algorithm for Grid Workflow

Grid architecture involves a large number of geographically distributed worker nodes connected together to achieve a level of performance. However, increasing the number of worker nodes does not always guarantee increased level of computing power. The resources involved in the system must be used such that all of the resources are utilized appropriately. The unequal demands and heterogeneity of Grid resources leads to the problem of job distribution. Algorithms try to map jobs on resources such that all the resources are utilized equally and makespan of the applications is reduced. Workload is the amount of work to be done by resource which can be heavy, light or moderate. Load balancing is sometime confused with load sharing and load leveling. Load sharing confirms that there is no idle node when there is highly loaded node in the Grid. This is the most basic level of load distribution which only checks if there is a load available on a resource or not, i.e., it is viewed as a binary set. Load balancing is the finest form of load distribution which tries to achieve strictly balanced distribution of load among all of the resources of the system. Load leveling is in the middle of the two extremes of load distribution which seeks to avoid congestion on any of the resources in the system.
Mostrar mais

10 Ler mais

Cost based Resource Provisioning Policy for Grids

Cost based Resource Provisioning Policy for Grids

The design of gateway that provision resources to dead- line applications relying on information given by current resource management services, may be complex and prone to scheduling decisions, that are far from optimal when providers and brokers use conflicting policies, the number of migrations can be high [13]. So if we will use same techniques with resources providers and user then automat- ically number of migrations and complexity of policies can be less and the performance will be high. In this work, a resource provisioning policy is designed in XML schema and validation and implementation of this policy has been done using Z-language.
Mostrar mais

4 Ler mais

Query Optimization Architecture for Data Grid Environment

Query Optimization Architecture for Data Grid Environment

The goal of query optimization is to determine a strategy for query execution that minimizes a cost function. The optimization phase includes: (i) selecting the best replication of relations, (ii) selecting the best nodes among the set of nodes discovered, (iii) the definition of the order of relational operations and the best access algorithm for each of them, and (iv) the placement of relational operators (also called resource allocation) for the execution of operations. The optimizer is based on a cost model consisting essentially of (i) a set of metrics (e.g. response time of a join) and (ii) a library. The metric (i) are calculated based on parameter hosts values (i.e. value of the CPU load, load value of I/O and value of the memory load of a node) and network (i.e. the value of the
Mostrar mais

8 Ler mais

Review Strategies and Analysis of Mobile Ad Hoc Network- Internet Integration Solutions

Review Strategies and Analysis of Mobile Ad Hoc Network- Internet Integration Solutions

The challenge in integration of MANET-Internet is to inform mobile nodes about available Internet gateways while making a minimal consumption of the scarce network resources. So, an efficient gateway discovery for Ad Hoc networks becomes one of the key elements to enable the use of hybrid Ad Hoc networks in future mobile and wireless networks. In order to be able to communicate with the Internet, each mobile node within the MANET must configure globally routable IP address [2]. This includes acquiring a temporary address once it enters the MANET, which allows only local communication, within the Ad Hoc network. Next it needs to discover and select one Internet gateway to use its prefix and form a globally routable address. This paper also tackles this fundamental problem and addresses the interworking between Ad Hoc networks and the Internet. Let us take a closer look at the interconnection of MANET and fixed network. Unlike the fully hierarchical addressing scheme used in the Internet, MANETs have a completely flat addressing model. In fact, Ad Hoc routing protocol such as AODV does not employ the concept of IP subnet. They assume that a node in a MANET may use any IP address provided that it is not duplicated. In fact, Ad Hoc routing protocols use host-based routes rather than network prefixes. Unlike in traditional IP networks, two neighboring nodes are not required to have addresses belonging to the same IP subnet to be able to directly communicate with each other. These differences with traditional IP networks create some interworking issues as indicated below [3].
Mostrar mais

11 Ler mais

Task Scheduling Algorithm to Reduce the Number of Processors using Merge Conditions

Task Scheduling Algorithm to Reduce the Number of Processors using Merge Conditions

The result in Table. 4 is affected by an iteration condition in Cluster_Merge() of Fig. 10 that stops the iteration if the next merge increases schedule length. Thus the reduction amount in the number of processors is not so remarkable. Let us change the iteration stop condition such that until there is no processor pair to be merged. That is, results by allocating task with higher out-degree prior and by using function Merge_Check()has better schedule length at the cost of smaller reduction in the number of processors. Table. V shows that applying function Cluster_Merge_MP() increases schedule length a little but reduces the combined cost much compared to function Cluster_Merge(). The number of processors is reduced to 11.95% ~ 21.06% while schedule length is increased to 3.05% ~ 6.89%. Let us define computation cost is a product of schedule length and the number of processors. The proposed merge algorithm generates schedules of which computation cost is 84.3% ~ 91.2% compared to the schedule by Cluster_Merge().
Mostrar mais

12 Ler mais

Scheduling in Desktop Grid Systems: Theoretical Evaluation of Policies & Frameworks

Scheduling in Desktop Grid Systems: Theoretical Evaluation of Policies & Frameworks

OurGrid is an open source middleware designed for peer- to-peer computational grids [46]. OurGrid enables the use of idle computing and storage resources over a grid. These resources are shared in such a way that who have contributed the most will get the most required. OurGrid provides a secure platform for the execution of parallel applications having independent tasks also called Bag-of-Tasks (BoT) applications. BoT examples may include parameter sweep simulations, rendering of images and many others. In OurGrid, each grid site corresponds to a peer in the system. The problem of free riders (people who are not contributing their resources but using resources of others) is resolved in OurGrid by using Network of Favours mechanism. This credit mechanism ensures that the computing node sharing its resources will be prioritized over a node that is not sharing the resources. OurGrid Community, a free-to-join cooperative grid is also maintained by OurGrid team.
Mostrar mais

9 Ler mais

Show all 10000 documents...