Adaptive divisible load scheduling strategies for workstation clusters with unknown network resources
Adaptive divisible load scheduling strategies for workstation clusters with unknown network resources
复制标题
DOI：
10.1109/tpds.2005.117
复制
发表时间：
2005-10
影响因子：
5.3
通讯作者：
Debasish Ghose;Hyoung-Joong Kim;Taehoon Kim
中科院分区：
文献类型：
--
作者：
Debasish Ghose;Hyoung-Joong Kim;Taehoon Kim
传统的可分负载调度算法试图在网络中存在通信延迟的情况下实现要在分布式计算系统中的处理器之间分配的大规模负载的最佳划分。然而，这些算法强烈地依赖于网络参数的先验知识的假设，并且不能处理关于这些参数的变化或缺乏信息。在本文中，我们提出了一种自适应的策略，估计网络参数值使用探测技术，并使用它们来获得最佳的负载划分。三种算法，基于相同的策略，提出的文件中，结合的能力，以科普未知的网络参数。给出了几个数值例子。最后，我们实现了一个实际的网络上的处理器节点使用MPI实现的自适应算法，并证明了自适应方法的可行性。
Conventional divisible load scheduling algorithms attempt to achieve optimal partitioning of massive loads to be distributed among processors in a distributed computing system in the presence of communication delays in the network. However, these algorithms depend strongly upon the assumption of prior knowledge of network parameters and cannot handle variations or lack of information about these parameters. In this paper, we present an adaptive strategy that estimates network parameter values using a probing technique and use them to obtain optimal load partitioning. Three algorithms, based on the same strategy, are presented in the paper, incorporating the ability to cope with unknown network parameters. Several illustrative numerical examples are given. Finally, we implement the adaptive algorithms on an actual network of processor nodes using MPI implementation and demonstrate the feasibility of the adaptive approach.