Dynamic Load Balancing for Programming Memory Efficient MapReduce Applications

碩士 === 中華大學 === 資訊工程學系碩士在職專班 === 100 === The term ‘cloud computing’ has been discussed widely in media such as in journals or magazines. Generally, ‘cloud’ refers to network. This is due to the reason that computer engineers usually use a cloud-like symbol network in their sketch diagrams. Therefo...

Full description

Bibliographic Details
Main Authors: Teng-Wei Hsu, 許登維
Other Authors: Ching-Hsien Hsu
Format: Others
Language:zh-TW
Published: 2012
Online Access:http://ndltd.ncl.edu.tw/handle/25065580294293524273
Description
Summary:碩士 === 中華大學 === 資訊工程學系碩士在職專班 === 100 === The term ‘cloud computing’ has been discussed widely in media such as in journals or magazines. Generally, ‘cloud’ refers to network. This is due to the reason that computer engineers usually use a cloud-like symbol network in their sketch diagrams. Therefore, cloud computing generally refers to network computing. For instance, any computing that is operated through network by computers, is a kind of cloud computing. The terminal services providers can use the cloud computing technique to process huge amount of information, reaching high efficiency network services. Under the framework of cloud computing, information has been disassembled into smaller pieces and assigned into different networked devices, and hence has been reassembled by computing. Therefore, disassemble, reassemble, as well as distribution of the information influence both the efficiencies of resources utilization and computing. This study investigates the best assignment of the information among the cloud computing devices, aiming to optimize the resource utilization and computing efficiencies. In this study, a dynamic distribution methodology is proposed to optimize the cloud computing. The proposed method reduces computing memory and enhances degree of parallelism during disassemble and reassemble processes. Particular for the heterogeneous information, such method enhances the efficiencies of resource utilization and computing by optimizing the distribution of information.