Fault-Tolerant Agreement-Related Problems in Distributed Cloud Computing Environment

View Statistics Email Alert RSS Feed

Information

Details

Project title

Code/計畫編號

NSC101-2221-E032-039

Translated Name/計畫中文名

分散式雲端網路環境下各種容錯協議相關問題之研究

Project Coordinator/計畫主持人

Chien-Fu Cheng

Funding Organization/主管機關

National Science and Technology Council

Department/Unit

Tamkang University Department of Computer Science and Information Engineering

Website

https://www.grb.gov.tw/search/planDetail?id=2632103

Year

2012

Start date/計畫起

01-08-2012

Expected Completion/計畫迄

31-07-2013

Bugetid/研究經費

594千元

ResearchField/研究領域

資訊工程--硬體工程

雲端運算(Cloud Computing)的概念在於透過無所不在的網路將一複雜運算，拆成許多較小的子程序，再經由遠端多部處理器所組成的雲端運算系統，執行一連串的搜尋、分析以及計算後，將處理完畢的結果合併，最後再透過網路回傳給使用者。因此，如何讓身處於雲端運算系統内的眾多處理器一起協同工作，並且有效的執行工作，嚴然成為一項非常重要的課題。然而，雲端運算系統可能因為處理器的損毁，使得系統無法正常運作，甚至讓整個系統癱瘓。因此，我們需要一個機制能夠讓參與運作的各種處理器即使在發生損毁或被攻擊時仍能夠協同工作。在本計畫中，我們將藉由探討分散式雲端網路環境上的拜占庭協議問題(Byzantine Agreement Problem, BA problem)、互動一致性問題(Interactive Consistency, IC)以及合議問題(Consensus Problem)來解決上述之協同問題。同時進一步將損毁情形區分為硬性損毁(Hardware fault)以及軟性損毁(Software fault)，藉由分析不同程度之損毁情況以提升整體之容錯能力。為了追求更好的通訊協定執行效率，我們將導入Early Stopping, (ES)的概念，設計出適用於分散式雲端網路環境下之ES規則，當收集到足夠的資訊時即停止訊息的交換，進入決議階段，以減少訊息所需的交換次數，提升通訊協定之執行效率。最後，在一個需要高可靠度的運算環境之下，僅僅達成容錯是不夠的，還需要能夠找出發生錯誤的損毁元件。因此，我們也將探討錯誤診斷協議問題(Fault Diagnosis Agreement Problem)於分散式雲端網路環境之中，藉此找出損毁的元件。在時程方面，本計畫預計以三年的時間來解決上述之問題。透過雲端運算，整合眾多電腦的資源使之協同合作以完成龐大的作業，是未來發展的必然趨勢。因此，對於雲端環境中的協議/合議相關問題有其被探討之必要性。 In recent years, cloud computing is becoming more and more ubiquitous. Cloud computing is to break down a complicated task into multiple sub-processes, use a cloud computing system comprising multiple remote processors to perform search, analysis, and computation of these processes over a ubiquitous network, and then deliver the combined result to the user. Thus, how to ensure collaboration and effective operation of multiple processors included in a cloud computing system is an important issue. As a cloud computing system may operate abnormally and even crash if any processor is damaged, we need a mechanism that can ensure robustness of the collaboration of all types of processors against dormant and malicious faults. In this project, we attempt to address the above problem through an investigation into Byzantine Agreement (BA) problem, Interactive Consistency (IC) problem, and Consensus Problem in the distributed cloud environment. We will divide faults into hardware-fault and software-fault and enhance the fault tolerance of the system through analysis of fault level. Further, we will introduce the Early Stopping (ES) concept to increase the efficiency of protocols. An ES rule suited for the distributed cloud environment will be developed. With this rule, the system can terminate the message exchange operation if enough messages have been collected and enter the decision-making stage earlier. Through reduction of message exchange rounds, this rule can increase the efficiency of the protocol. Finally, a cloud computing system should have the ability to identify faulty components in addition to fault tolerance. Hence, we will also discuss Fault Diagnosis Agreement (FDA) problem in the distributed cloud environment and develop a mechanism that can help the system identify faulty components. This project is expected to take three years to complete. As utilizing multiple computer resources to perform complicated tasks via cloud computing is a foreseeable trend, we see the necessity of investigating agreement problems in the cloud environment.

Keyword(s)

雲端運算
拜占庭協議問題
互動一致性問題
合議問題
即時協議
錯誤診斷問題
硬性毁損
軟性毁損
cloud computing
byzantine agreement problem
interactive consistency
consensus problem
early stopping agreement
fault diagnosis problem
hardware-fault
software-fault

DSpace CRIS

Fault-Tolerant Agreement-Related Problems in Distributed Cloud Computing Environment

Details

Description