Web analytics has become an important business and market research tool for measuring and improving the effectiveness of a website. Web analytics requires to log in the database, information like the number of visitors visiting a web page or country of the visitor, some other information related to the visitor or web page. Considering the fact that some web pages attract huge traffic in relatively short time, either because it has gone viral or due to marketing or any other reason. As the number of page views increase exponentially, the rate of requests coming to web analysis server and entries into their database can become significantly high. In such cases, server might not sustain such load and which will consequently affect the web analysis of server's performance. This paper proposes a request control system which dynamically updates the amount of requests coming from a web page under analysis, based on the number of visitors visiting the web page, and still is able to maintain the accurate count of visitors visiting web page with error less than 1. 1%. This paper presents the performance measurements for the proposed system as well as analyzes the efficiency and limitations of this approach.

Web analytics Content Delivery Network Redis Internet Marketing