云計(jì)算環(huán)境下煤礦應(yīng)急管理關(guān)鍵技術(shù)研究

1、 論文題目：云計(jì)算環(huán)境下煤礦應(yīng)急管理關(guān)鍵技術(shù)研究 專 業(yè)：安全技術(shù)及工程 博 士 生：馬 莉 （簽名） 指導(dǎo)教師：李樹剛 （簽名） 摘 要 隨著煤礦應(yīng)急管理信息化建設(shè)的不斷深入， 煤礦應(yīng)急管理信息系統(tǒng)也將隨著云計(jì)算技術(shù)朝智能化、服務(wù)化方向發(fā)展。本文采用云計(jì)算―按需服務(wù)‖的計(jì)算

2、思想和面向服務(wù)的架構(gòu)，以云計(jì)算環(huán)境下的煤礦應(yīng)急管理信息技術(shù)為研究對象，以云服務(wù)的方式為煤礦用戶提供應(yīng)急信息資源服務(wù)，從煤礦應(yīng)急管理的新模式、云作業(yè)調(diào)度、云存儲(chǔ)副本布局、云服務(wù)語義描述、海量數(shù)據(jù)管理五個(gè)方面來開展研究。主要研究內(nèi)容如下： 研究了云計(jì)算環(huán)境下的煤礦應(yīng)急管理新模式。 結(jié)合應(yīng)急管理理論、 ―情景-應(yīng)對‖范式和權(quán)變理論， 構(gòu)建煤礦安全事故動(dòng)態(tài)應(yīng)急管理模式； 基于面向服務(wù)構(gòu)架和云計(jì)算的思想，提出云計(jì)算環(huán)境下煤礦應(yīng)急管理的新模式——基

3、于SOA的煤礦應(yīng)急云CECloud，并提出基于―情景-應(yīng)對-權(quán)變‖模式的煤礦應(yīng)急云運(yùn)作機(jī)制；結(jié)合信息服務(wù)理論，提出煤礦應(yīng)急信息資源云及云服務(wù)的概念。 研究了煤礦應(yīng)急云的動(dòng)態(tài)反饋?zhàn)鳂I(yè)調(diào)度算法。 針對現(xiàn)有Hadoop作業(yè)調(diào)度算法在多用戶、異構(gòu)環(huán)境下不具備反饋機(jī)制的問題，結(jié)合煤礦應(yīng)急云的用戶作業(yè)特征提出一種云計(jì)算環(huán)境下具備反饋機(jī)制的動(dòng)態(tài)作業(yè)調(diào)度算法(CECloud-MMS Scheduler)。該算法引入排隊(duì)論模型，采用單隊(duì)列多資源池服務(wù)窗口

4、的設(shè)計(jì)思路，將所有作業(yè)統(tǒng)一提交到一個(gè)支持優(yōu)先級的排隊(duì)隊(duì)列， 作業(yè)分發(fā)控制模塊選擇優(yōu)先級最高的作業(yè)分發(fā)到空閑的資源池窗口執(zhí)行；煤礦應(yīng)急云通過自身的心跳機(jī)制將作業(yè)運(yùn)行的初始化時(shí)間、運(yùn)行時(shí)間等信息傳遞給參數(shù)統(tǒng)計(jì)模塊進(jìn)行統(tǒng)計(jì)， 將獲得的平均到達(dá)率和平均服務(wù)率這兩個(gè)核心參數(shù)的實(shí)際值傳遞給反饋機(jī)制模塊， 根據(jù)調(diào)度算法模型計(jì)算出平均逗留時(shí)間和平均隊(duì)長的理論值并與實(shí)際值進(jìn)行對比， 當(dāng)差值大于閾值時(shí)對該調(diào)度算法的核心參數(shù)進(jìn)行適當(dāng)調(diào)節(jié)使差值收斂于閾值， 則

5、將具有較大平均逗留時(shí)間和平均隊(duì)長的作業(yè)調(diào)度到有槽位數(shù)的空閑資源池服務(wù)窗口執(zhí)行。實(shí)驗(yàn)結(jié)果表明，與經(jīng)典算法相比，CECloud-MMS Scheduler具有較高的調(diào)度效率和負(fù)載平衡能力。 研究了煤礦應(yīng)急云存儲(chǔ)低代價(jià)副本布局技術(shù)。 提出云計(jì)算環(huán)境下的煤礦應(yīng)急管理數(shù)據(jù)存儲(chǔ)方式及煤礦應(yīng)急管理信息系統(tǒng)應(yīng)用體系， 提出基于低訪問代價(jià)的煤礦應(yīng)急云存儲(chǔ)副本布局策略，采用TOPSIS分析方法進(jìn)行實(shí)例計(jì)算，將副本分布在煤礦應(yīng)急云平臺(tái)中訪問代價(jià)最低的2個(gè)服務(wù)

6、器節(jié)點(diǎn)上，在節(jié)點(diǎn)存儲(chǔ)負(fù)載、節(jié)點(diǎn)網(wǎng)絡(luò)帶寬、節(jié)點(diǎn)響應(yīng)率、節(jié)Subject ：Study on Key Technique of Coalmine Emergency Management in Cloud Computing Specialty ：Safety Technology and Engineering Name ：Ma Li （Signature） In

7、structor：Li Shugang （Signature） ABSTRACT With the development of informatization construction of coal mine emergency management, the emergency information system will toward intelligent, se

8、rvice-oriented direction. The paper adopts ―Service on demand‖ computational thinking of cloud computing and service-oriented architecture to study emergency technique in cloud computing and provide emergency information

9、 resource service with type of cloud service for coalmine users. Five aspects of new mode, cloud job scheduling, cloud storage copy layout, description of cloud service and massive data processing are studied in coalmine

10、 emergency management. The main research contents are as follows. Study on coalmine emergency new mode in cloud computing. Conbined with emergency management theory, ―scenario-respond‖ paradigm and contingency theory, th

11、e coalmine accidents dynamic emergency management mode is constructed. Based on ―Service on demand‖ computational thinking of cloud computing, a new mode of coal mine emergency management is proposed with ―Coalmine Emerg

12、ency Cloud (CECloud) based on SOA‖ as well as operating mechanism of ―scenario-respond-contingency‖mode. With information services theory, the concept of coalmine emergency information resources cloud and cloud service a

13、re put forward. Study on dynamic and feedback job scheduling algorithm of coalmine emergency cloud. Aiming at defectiveness in the feedback mechanism of current Hadoop scheduling algorithms in a multi-user and heterogene

14、ous environment, a dynamic job scheduling algorithm CECloud-MMS Scheduler with feedback mechanism is proposed. In this algorithm, a kind of Queues Theory model and design of single queue and multiple resource pool servic