The Durability of the “Bow-Tie Structure of the Web”
SOURCE:
https://www.i-programmer.info/news/82-heritage/10681-google-researchers-win-award-for-classic-paper.html
In April of 2017, the researchers whose work revealed the “bow-tie” structure of the web were awarded the third Seoul Test of Time Award. This award is given to the authors of a previous World Wide Web Conference paper that has demonstrated significant impact in the years since its publication. At the time of the award ceremony, the paper had accumulated over 3,500 citations in 17 years. The paper made a major contribution to our understanding of the structure of the Internet. Previous research suggested that the Web was almost fully connected, however this paper proposed a much more elaborate structure with a “bow-tie” shape. This suggested that the probability that any two web pages can be reached from one another was lower than previously predicted – just under ¼.
We discussed the “bow-tie” in class and analyzed what characteristic classes of Web pages this structure describes. There is the strongly connected component (SCC) at the center, in which every page is reachable from any other page. There are the “in” and “out” clusters, which have unidirectional paths to or from the SCC. There are the tendrils that attach to the clusters and tubes that connect the clusters. Lastly, there are the disconnected components, which are isolated from the rest of the graph.
This was one of the first large-scale studies of the dynamics of the Web. Without the findings from this paper, search engines would not be able to continuously improve their search results. I think it is truly incredible that though the Web has expanded significantly since this paper’s publication in 2000, the “bow-tie” structure still holds true and continues to enhance the quality of search results. A revised model was created in 2014, where the SCC now had 1.8 billion nodes. That number would be even greater if a new model was created today.