Internet statistical measurement - sampling through overbiasing

In this paper, we propose an IP address-based approach to track the growth of the Web for making a reliable and accurate estimate of the IP address usage. Our method uses a proper sampling of the entire IP address space, resulting in a significant reduction of effort to obtain an accurate estimate. A thorough analysis of the proposed improved Importance Sampling approach is presented. Based on the measurement of the number of active Web servers over the past years, an exponential growth of the Web is observed and modeled.