一种集成用户画像与内容的服务重定向方法

Approach to Service Redirection Integrated with User Portrait and Content

随着互联网的发展,网站应用的规模正不断扩大,常规的垂直应用架构已慢慢无法应对这样的场景,而应用的大规模服务化则应运而生.大规模服务框架通常会建立一个服务注册中心,动态的注册和发现服务,使服务的位置透明.消费方通过获取服务提供方的地址列表,实现软负载均衡和Failover.而随着服务的累积,在候选服务集合越来越大的情况下,加快重定向的响应速度是其中的一个关键问题.本文旨在通过cookies跨域采集用户行为信息并对用户分群画像,用LDA分析网页内容并建立主题模型,进而提出一种基于用户画像与内容的服务重定向方法.该方法基于人群特征与内容修剪候选服务,可以大大减少搜索空间,降低计算量,以提高响应速度.实验结果验证了本文方法的有效性.

With the development of Internet,the scale of Web application continuously extends, which leads to that conventional vertical application architecture has been unable to cope with large-scale services. Large-scale service framework usually establishes a service registry center,which can dynamically register discover service with service location transparency. Consumer achieves soft load balancing and failover by getting the provider address list. With the accumulation of services ,it is a key issue to accelerate the response speed of redirection according to more and more huger candidate services. This paper focuses on cross-domain collecting user behavior information through cookies ,draws group portrait for users and establishes LDA model. Furthermore,we propose a service redirection approach based on user portrait and content, which can prune candidate services, reduce search space greatly and improve the response speed. To validate our approach, large-scale experiments are conducted. The results show the effectiveness of the approach in this paper.