#### DMCA

## Bag-of-Tasks Scheduling under Budget Constraints (2010)

Venue: | In 2nd IEEE International Conference on Cloud Computing Technology and Science (CloudCom 2010 |

Citations: | 22 - 5 self |

### Citations

731 |
Knapsack Problems: Algorithms and Computer Implementation , pp 77-80
- Martello, Toth
- 1990
(Show Context)
Citation Context ...astest combination of machines whose total cost per ATU is within a given limit. Though BKP is NP-complete, it can be solved in pseudopolynomial time either by expressing it as a 0-1 Knapsack Problem =-=[11]-=- or by using a specialized algorithm [12]. Since both the number of machine types as well as the number of machines of each type are small (compared to the number of jobs), the input of the reformulat... |

538 | Distributed computing in practice: the condor experience. Concurrency and Computation: Practice and Experience
- Thain, Tannenbaum, et al.
- 2005
(Show Context)
Citation Context ...aring to a cost-oblivious round-robin scheduler. 1 Introduction In computational science, parameter sweep or bag of tasks applications are as dominant as computationally demanding. The classic Condor =-=[16]-=- system is in widespread use to deploy as many application tasks as possible on otherwise under utilized computers, coining the term of High Throughput Computing, utilizing networks of idle workstatio... |

241 | Heuristic algorithms for scheduling independent tasks on nonidentical processors
- Ibarra, Kim
- 1977
(Show Context)
Citation Context ...us of research done in [3, 9]. This is accompanied by response time minimization at task level in [17]. These scheduling algorithms are compared against traditional algorithms such as MinMin, Max-Min =-=[5]-=- and Sufferage [10]. A mixture of robustness optimization, while satisfying makespan and price constraints is presented in [15]. It assumes a fixed, one-time cost per machine type. Robustness optimiza... |

169 | Dynamic Matching and Scheduling of a Class of Independent Tasks onto Heterogeneous Computing Systems
- Maheswaran, Siegel, et al.
- 1999
(Show Context)
Citation Context ...e in [3, 9]. This is accompanied by response time minimization at task level in [17]. These scheduling algorithms are compared against traditional algorithms such as MinMin, Max-Min [5] and Sufferage =-=[10]-=-. A mixture of robustness optimization, while satisfying makespan and price constraints is presented in [15]. It assumes a fixed, one-time cost per machine type. Robustness optimization is the main fo... |

54 | The performance of bags-oftasks in large-scale distributed systems
- Iosup, Sonmez, et al.
- 2008
(Show Context)
Citation Context ...o not allow queueing delays due to competing requests. We evaluated BaTS using a workload of relatively medium size, where the assumptions made by the statistical device behind BaTS hold. Recent work =-=[6]-=- on the properties of bags-of-tasks has shown that in many cases the intra-BoT distribution of execution times follows a normal distribution. Accordingly, we have generated a workload of 1000 tasks wh... |

52 |
Introduction to Statistical Inference
- KEEPING
- 1962
(Show Context)
Citation Context ... of data points; one set per cluster. The size n of the sample set can be computed with respect to a certain confidence level, based on the canonical statistical formula for sampling with replacement =-=[7]-=-: ⌈ n = N ∗ z 2 α z 2 α + 2 ∗ (N − 1)∆ 2 ⌉ , where n is the sample size, N is the size of the bag, ∆ ∈ {0.10, 0.15, 0.20, 0.25} are typical values for the error level, α ∈ (0, 1) is the confidence lev... |

34 |
Heuristic Scheduling for Bag-of-Tasks Applications in Combination with QoS
- Weng, Lu
- 2005
(Show Context)
Citation Context ...nd the respective algorithms. The assumptions on task characteristics involve the existence of prior knowledge on arrival rate, execution time, or deadline for each task in the bag. Work presented in =-=[9, 14, 15, 17]-=- assumes a-priori known task execution times. One relaxation is found in [4] where all tasks are assumed to be in the same complexity class and there is a calibration step to determine execution time ... |

26 |
Robust Static Allocation of Resources for Independent Tasks under Makespan and Dollar Cost Constraints
- Sugavanam
(Show Context)
Citation Context ...nd the respective algorithms. The assumptions on task characteristics involve the existence of prior knowledge on arrival rate, execution time, or deadline for each task in the bag. Work presented in =-=[9, 14, 15, 17]-=- assumes a-priori known task execution times. One relaxation is found in [4] where all tasks are assumed to be in the same complexity class and there is a calibration step to determine execution time ... |

25 | Running bag-of-tasks applications on computational grids: The MyGrid approach
- Cirne, Paranhos, et al.
- 2003
(Show Context)
Citation Context ...etect it while executing the bag of tasks. Work presented in [18] addresses homogenous grid resources and is extended to heterogenous systems in [19]. Completely heterogenous systems are addressed by =-=[3, 13, 15, 17]-=-. Our system is composed of several heterogenous sets of homogenous machines, which fits perfectly a composite of scientific grids and cloud systems. Makespan minimization is the main focus of researc... |

22 |
Zomaya A Y. Practical scheduling of Bag-of-Tasks applications on grids with dynamic resilience
- Lee
(Show Context)
Citation Context ...nd the respective algorithms. The assumptions on task characteristics involve the existence of prior knowledge on arrival rate, execution time, or deadline for each task in the bag. Work presented in =-=[9, 14, 15, 17]-=- assumes a-priori known task execution times. One relaxation is found in [4] where all tasks are assumed to be in the same complexity class and there is a calibration step to determine execution time ... |

16 |
Real-World Distributed Computing with Ibis
- Bal, Maassen, et al.
(Show Context)
Citation Context ...ous sections. If a budget constraint violation is signaled, BaTS tries to compute a new configuration. 3 Performance Evaluation We have implemented a Java-based BaTS prototype using our Ibis platform =-=[2]-=-. The prototype consists of two parts: the master, which can run on any desktop-like machine, and the worker which is deployed on cloud machines. The master component implements the core of BaTS, whil... |

12 | Measuring the robustness of resource allocations in a stochastic dynamic environment
- Smith, Briceño, et al.
- 2007
(Show Context)
Citation Context ...nd in [4] where all tasks are assumed to be in the same complexity class and there is a calibration step to determine execution time estimates per machine type. The assumptions are further relaxed in =-=[13]-=- to relative complexity classes of tasks, though all the classes are supposed to be known in advance. BaTS, in contrast, only assumes that some form of runtime distribution exists, and uses stochastic... |

11 |
Self-adaptive skeletal task farm for computational grids. Parallel Comput. (2006
- González-Vélez
(Show Context)
Citation Context ...nce of prior knowledge on arrival rate, execution time, or deadline for each task in the bag. Work presented in [9, 14, 15, 17] assumes a-priori known task execution times. One relaxation is found in =-=[4]-=- where all tasks are assumed to be in the same complexity class and there is a calibration step to determine execution time estimates per machine type. The assumptions are further relaxed in [13] to r... |

10 |
A Minimal Algorithm for the Bounded Knapsack Problem
- Pisinger
(Show Context)
Citation Context ...l cost per ATU is within a given limit. Though BKP is NP-complete, it can be solved in pseudopolynomial time either by expressing it as a 0-1 Knapsack Problem [11] or by using a specialized algorithm =-=[12]-=-. Since both the number of machine types as well as the number of machines of each type are small (compared to the number of jobs), the input of the reformulated BKP can be considered small in its len... |

7 | An adaptive middleware for supporting time-critical event response
- Zhu, Agrawal
- 2008
(Show Context)
Citation Context ...sed to be known in advance. BaTS, in contrast, only assumes that some form of runtime distribution exists, and uses stochastic methods to detect it while executing the bag of tasks. Work presented in =-=[18]-=- addresses homogenous grid resources and is extended to heterogenous systems in [19]. Completely heterogenous systems are addressed by [3, 13, 15, 17]. Our system is composed of several heterogenous s... |

6 | A resource allocation approach for supporting time-critical event handling in grid environments
- Zhu, Agrawal
- 2009
(Show Context)
Citation Context ...e distribution exists, and uses stochastic methods to detect it while executing the bag of tasks. Work presented in [18] addresses homogenous grid resources and is extended to heterogenous systems in =-=[19]-=-. Completely heterogenous systems are addressed by [3, 13, 15, 17]. Our system is composed of several heterogenous sets of homogenous machines, which fits perfectly a composite of scientific grids and... |

4 | Robustprocessor allocation for independent tasks when dollar cost for processors is a constraint
- Sugavanam, Maciejewski, et al.
- 2005
(Show Context)
Citation Context |