本文選題：分布式服務(wù)框架 + 服務(wù)重定向��； 參考：《小型微型計算機(jī)系統(tǒng)》2017年12期

【摘要】：隨著互聯(lián)網(wǎng)的發(fā)展,網(wǎng)站應(yīng)用的規(guī)模正不斷擴(kuò)大,常規(guī)的垂直應(yīng)用架構(gòu)已慢慢無法應(yīng)對這樣的場景,而應(yīng)用的大規(guī)模服務(wù)化則應(yīng)運(yùn)而生.大規(guī)模服務(wù)框架通常會建立一個服務(wù)注冊中心,動態(tài)的注冊和發(fā)現(xiàn)服務(wù),使服務(wù)的位置透明.消費(fèi)方通過獲取服務(wù)提供方的地址列表,實(shí)現(xiàn)軟負(fù)載均衡和Failover.而隨著服務(wù)的累積,在候選服務(wù)集合越來越大的情況下,加快重定向的響應(yīng)速度是其中的一個關(guān)鍵問題.本文旨在通過cookies跨域采集用戶行為信息并對用戶分群畫像,用LDA分析網(wǎng)頁內(nèi)容并建立主題模型,進(jìn)而提出一種基于用戶畫像與內(nèi)容的服務(wù)重定向方法.該方法基于人群特征與內(nèi)容修剪候選服務(wù),可以大大減少搜索空間,降低計算量,以提高響應(yīng)速度.實(shí)驗(yàn)結(jié)果驗(yàn)證了本文方法的有效性.
[Abstract]:With the development of the Internet, the scale of website application is expanding. The conventional vertical application architecture has been unable to cope with this kind of scenario, and the large-scale service of application has come into being. Large-scale service frameworks usually establish a service registry to dynamically register and discover services, making the location of services transparent. The consumer achieves soft load balancing and failure over by obtaining the address list of the service provider. With the accumulation of service, it is a key problem to accelerate the response speed of redirect when the candidate service set becomes larger and larger. The purpose of this paper is to collect user behavior information through cookies across domains, divide users into groups, analyze web content with LDA and set up theme model, and then propose a service redirection method based on user portrait and content. Based on crowd feature and content pruning candidate services, this method can greatly reduce the search space, reduce the amount of computation, and improve the response speed. The experimental results show that the proposed method is effective.
【作者單位】： 湖北文理學(xué)院;武漢大學(xué)國際軟件學(xué)院;
【基金】：國家重點(diǎn)研發(fā)計劃項目(2016YFB0800400)資助 國家自然科學(xué)基金項目(61572371)資助 湖北省自然科學(xué)基金面上項目(2016CFB406)資助 湖北文理學(xué)院教師科研能力培育基金項目(2016ZK004)資助
【分類號】：TP393.092

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 陳其杉;方勇;李岳t，

本文編號：1889040