Re-Examining the Complexity of Popular Websites

Abstract

During the past decade, the Web has become increasingly more popular and thus more important for delivery of content and services over the Internet. At the same time, the number of requested objects, their size and delivery mechanisms for popular websites have become more complex. This in turn has various implications including the impact on page loading time that directly affects the experience of visiting users. Therefore, it is important to capture and characterize the complexity of popular web pages. An earlier study by Butkiewicz et al. have characterized the complexity of 1700 popular pages in 2011. In this study, we adopt the methodology proposed by Butkiewicz et al., develop the required tools and conduct a detailed measurement study to re-assess the complexity of 2000 popular web pages and present any observed trends in their complexity characteristics over the past four years. Our results show that the number of requested objects and contacted servers for each website has significantly increased. But a growing number of contacted servers are associated with third parties. Despite these changes, the page loading time remains rather unchanged and it is primarily affected by the same key parameters. Overall, our results sheds a useful light on trends in web site complexity and motivates a range of issues to be explored.