Performance of Dynamic Web Page Generation for Database-driven Web Sites

With the rapid growth of e-business, Web sites are increasingly relying on dynamic content generation applications to provide dynamic interaction and personalized experiences to the Web sites users. To generate such dynamic Web page for each user request requires heavy dynamic script execution due to customize and personalize information. This dynamic script execution puts heavy load on the Web server. Again database server experiences extra overheads due to increase of database-driven e-commerce applications and their frequently update. These extra loads on Web server and database server lead to increase response time. The overall response time is made up of two types of latency, namely network latency and server-side latency. In this paper we are mainly concerned about server-side latency. We have proposed an approach which integrates prediction based dynamic Web page pre-generation with PFC and DBC caching. Our integrated approach reduces dynamic Web page generation time by achieving the eventual benefits through middle-tier database cache (DBC) and HTML fragment cache(PFC) and the immediate benefit through dynamic Web page pre-generation. Response time is reduced by storing database tables in the middle-tier cache and by sharing common fragments among multiple pages stored in the HTML fragment cache. The reduction of response time latency significantly improves the performance and scalability

1. Introduction

With the rapid growth of dynamics web sites and e-Business, the web content becoming more dynamic coincides with enhanced personalization. By using dynamic page generation technology, a web site can generate pages at run time, significantly increasing its flexibility in customizing page content and thereby providing the users more personal and rich web experience. Generating personalized content is computationally expensive and further reduces the effectiveness of proxy caches. These dynamic and personalized page contents severely reduce web and application server performance and scalability because of the additional load placed on the web and application server. The delay for this load includes: (1) computational delay (such as from query processing), (2) interaction bottlenecks (such as from waiting for a DBMS connection), (3) cross-tier communication (TCP/IP stack traversal, for example). (4) Content conversion (XML-to-HTML transformation).