基于分布式lamp架構(gòu)的web服務(wù)性能分析
發(fā)布時(shí)間：2019-05-08 05:15
【摘要】：Web的特性有交互性、廣泛性、易用性、實(shí)時(shí)性等，當(dāng)Web出現(xiàn)之后，使互聯(lián)網(wǎng)上的網(wǎng)頁(yè)數(shù)量以指數(shù)級(jí)飛速增長(zhǎng)。隨著網(wǎng)絡(luò)上流量的劇增，使一些Web站點(diǎn)負(fù)荷超載而變得反應(yīng)緩慢甚至土崩瓦解。有很多網(wǎng)站都是在發(fā)展初期，由于無(wú)法承受巨大的壓力導(dǎo)致崩潰。為了保證正常的web服務(wù)，必須搭建一套高效、穩(wěn)定、健壯、7*24小時(shí)不間斷的系統(tǒng)。基于分布式lamp架構(gòu)對(duì)web的服務(wù)性能進(jìn)行分析，可以找到Web服務(wù)組合型系統(tǒng)的性能的較優(yōu)配置，而且結(jié)果比較精確。Web服務(wù)正成為不同企業(yè)的分布式應(yīng)用之間通信的最重要的技術(shù)�，F(xiàn)階段Web服務(wù)已經(jīng)作為關(guān)鍵技術(shù)基礎(chǔ)服務(wù)于眾多的Web的分布式應(yīng)用(如電子政務(wù)、電子商務(wù)等)。 雖然現(xiàn)在關(guān)于服務(wù)器性能的優(yōu)化已經(jīng)有了成熟的發(fā)展，但是現(xiàn)在還存在多種問(wèn)題，比如隨著互聯(lián)網(wǎng)用戶群的膨脹，用戶數(shù)量的增加，服務(wù)器不能單靠增加服務(wù)器來(lái)保持穩(wěn)定的運(yùn)營(yíng)。從數(shù)據(jù)的存儲(chǔ)和分發(fā)、多級(jí)緩存、負(fù)載均衡到前段的CSS和JS，都是經(jīng)過(guò)完美的優(yōu)化和處理，來(lái)解決日益增加的負(fù)載。當(dāng)然，還有很多敏捷項(xiàng)目，他們用戶群小，對(duì)系統(tǒng)的壓力小，開(kāi)發(fā)速度快，敏捷的特性使得項(xiàng)目和產(chǎn)品多變，，這使得我們可以更快更好的滿足用戶的需求，當(dāng)然這也導(dǎo)致了系統(tǒng)在開(kāi)發(fā)的過(guò)程中，難免變得脆弱，越來(lái)越大的系統(tǒng)，導(dǎo)致服務(wù)器負(fù)載增加，甚至出現(xiàn)了1核CPU使用100%的情況。對(duì)于這些問(wèn)題，我們不斷的總結(jié)、反思，著手優(yōu)化，不斷升級(jí)，不斷試驗(yàn)，用新的技術(shù)創(chuàng)新去解決這些問(wèn)題。 綜上，本文通過(guò)apache服務(wù)器負(fù)載均衡，php框架性能分析及使用，mysql的優(yōu)化，Rpc構(gòu)建等幾個(gè)方面來(lái)對(duì)服務(wù)器性能進(jìn)行優(yōu)化，基于效率、代碼維護(hù)成本、對(duì)現(xiàn)有系統(tǒng)的影響程度、學(xué)習(xí)成本方面對(duì)調(diào)優(yōu)試驗(yàn)結(jié)果數(shù)據(jù)進(jìn)行分析，得出結(jié)論：高性能web服務(wù)的搭建，需要對(duì)以上模塊的合理利用，通過(guò)負(fù)載均衡分散用戶請(qǐng)求，memcached，分布式存儲(chǔ)session信息，合理的使用框架和表結(jié)構(gòu)。
[Abstract]:The characteristics of Web are interactive, extensive, easy to use, real-time, and so on. After the appearance of Web, the number of web pages on the Internet increases exponentially. With the rapid increase of traffic on the network, some Web sites are overloaded and become slow to respond and even collapse. There are many sites in the early stages of development, due to the unbearable pressure caused by collapse. In order to ensure a normal web service, we must build an efficient, stable, robust, 7-hour 24-hour system. By analyzing the service performance of web based on the distributed lamp architecture, we can find the optimal configuration of the performance of the Web services composite system. Web services are becoming the most important communication technology among distributed applications in different enterprises. At present, as the key technology foundation, Web service has been used in many distributed applications of Web (e-government, e-commerce, etc.). Although there has been a mature development in the optimization of server performance, there are still many problems, such as the expansion of the Internet user base and the increase in the number of users. Servers cannot be kept stable by simply adding them. From the storage and distribution of data, multi-level caching, load balancing to the front of the CSS and JS, are perfectly optimized and processed to solve the increasing load. Of course, there are a lot of agile projects, they have a small user base, less pressure on the system, rapid development, agile features make projects and products changeable, which allows us to meet the needs of users faster and better. Of course, this also led to the development of the system, inevitably become fragile, more and more large systems, resulting in an increase in server load, and even the use of 100% of the one-core CPU situation. For these problems, we continue to summarize, reflect, proceed to optimize, constantly upgrade, continue to experiment, with new technological innovation to solve these problems. In summary, this paper optimizes the server performance through apache server load balancing, php framework performance analysis and use, mysql optimization, Rpc construction and so on, based on efficiency, code maintenance cost, and the degree of impact on the existing system. From the aspect of learning cost, it is concluded that the construction of high-performance web service requires the rational utilization of the above modules, distributing user requests through load balancing, and distributed storage of session information by memcached, and it is concluded that the construction of high-performance session services requires a reasonable utilization of the above modules. Rational use of framework and table structure.
【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP393.09
本文編號(hào)：2471645
[Abstract]:The characteristics of Web are interactive, extensive, easy to use, real-time, and so on. After the appearance of Web, the number of web pages on the Internet increases exponentially. With the rapid increase of traffic on the network, some Web sites are overloaded and become slow to respond and even collapse. There are many sites in the early stages of development, due to the unbearable pressure caused by collapse. In order to ensure a normal web service, we must build an efficient, stable, robust, 7-hour 24-hour system. By analyzing the service performance of web based on the distributed lamp architecture, we can find the optimal configuration of the performance of the Web services composite system. Web services are becoming the most important communication technology among distributed applications in different enterprises. At present, as the key technology foundation, Web service has been used in many distributed applications of Web (e-government, e-commerce, etc.). Although there has been a mature development in the optimization of server performance, there are still many problems, such as the expansion of the Internet user base and the increase in the number of users. Servers cannot be kept stable by simply adding them. From the storage and distribution of data, multi-level caching, load balancing to the front of the CSS and JS, are perfectly optimized and processed to solve the increasing load. Of course, there are a lot of agile projects, they have a small user base, less pressure on the system, rapid development, agile features make projects and products changeable, which allows us to meet the needs of users faster and better. Of course, this also led to the development of the system, inevitably become fragile, more and more large systems, resulting in an increase in server load, and even the use of 100% of the one-core CPU situation. For these problems, we continue to summarize, reflect, proceed to optimize, constantly upgrade, continue to experiment, with new technological innovation to solve these problems. In summary, this paper optimizes the server performance through apache server load balancing, php framework performance analysis and use, mysql optimization, Rpc construction and so on, based on efficiency, code maintenance cost, and the degree of impact on the existing system. From the aspect of learning cost, it is concluded that the construction of high-performance web service requires the rational utilization of the above modules, distributing user requests through load balancing, and distributed storage of session information by memcached, and it is concluded that the construction of high-performance session services requires a reasonable utilization of the above modules. Rational use of framework and table structure.
【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP393.09
【參考文獻(xiàn)】
相關(guān)期刊論文 前1條
1 馬琳,羅鐵堅(jiān),宋進(jìn)亮,葉世偉;Web性能測(cè)試與預(yù)測(cè)[J];中國(guó)科學(xué)院研究生院學(xué)報(bào);2005年04期
本文編號(hào)：2471645
本文鏈接：http://www.wukwdryxk.cn/guanlilunwen/ydhl/2471645.html
最近更新
教材專著
