Web pages include extraneous material that may be viewed as undesirable by a user. Increasingly many Web sites also require users to register to access either all or portions of the site. Such tension between content owners and users has resulted in a "cat and mouse" game between content provided and how users access it. We carried out a measurement-based study to understand the nature of extraneous content and its impact on performance as perceived by users. We characterize how this content is distributed and the effectiveness of blocking mechanisms to stop it as well as countermeasures taken by content owners to negate such mechanisms. We also examine sites that require some form of registration to control access and the attempts made to circumvent it. Results from our study show that extraneous content exists on a majority of popular pages and that a 25-30% reduction in downloaded objects and bytes with corresponding latency reduction can be attained by blocking such cont...
Balachander Krishnamurthy, Craig E. Wills