The massive popularity of search engines lies in its ability to turn up results for search queries very quickly and accurately (University of
Google's Page Rank algorithm, which was named in part after Larry Page, analyses recursive links between sites. This works by assigning every page, its own Page Rank which is a specified link to the page, and how important the linking pages are. This helps to determine the sites that are of actual importance, and enables Google to deliver results that are as accurate as possible. Sites on the web are gathered by the Gogebic, a Web Crawler which trawls the internet going through links. Google then creates a digital copy or a cache of pages that are new or actively tries to update the cache, removing links that are no longer working and updating pages that have been changed. All this takes place on Google server (Price, G., (2001)). The caches purpose is to allow dead links to be viewed even though the source is down for whatever reason. The pages do not stay in the Google cache, although a useful feature has been the source of some legal disputes (Olsen, S., 2003).
Google’s ability to handle such a huge volume of traffic lies in its unique computer architecture, which distributes a search query over many servers operating in small networks. Queries are distributed over the networks depending on the users physical proximity to that network. According to a consensus, it is believed that Google has an estimated 100,000 servers, although a somewhat more accurate estimate puts the number of servers at between 63,272 machines to 79,112 machines. On the higher end, Google would have whopping 316,448 GHz of computing power, 158,224 GB of RAM and 6,180 Tb of hard disk space (Levy, S., 2004, Tristan, L., 2004).
------------------------------------------------------------------------------------------------
1 comments:
Post a Comment