David C. Wheeler and Morton E. O'Kelly of Ohio State University conducted an experiment designed to measure internet performance and assess both telecommunication access and success. Furthermore, the two observed which U.S. cities are the most efficient in both criteria.
The Internet was originally designed to connect researchers at several national supercomputers and provide a fast medium to share information. The original project was named The National Science Foundation or NSFNET. This backbone was completed in 1988. As the need for global telecommunications and commercial traffic grew, NSFNET gave way to the Commercial Internet Exchange (CIX) and ultimately, Network Access Points (NAP) and Metropolitan Area Ethernet (MAE). From the mid to late 90s' the Internet grew exponentially as its need for businesses purposes and private use increased. The Internet's backbones NAP and MAE serve as networks where Internet Service Providers (ISPs) can connect to exchange information. To date, this is the type of Internet used by businesses and the public.
Many attempts have been made to measure the Internet's performance. Studies such as the Matrix Information and Directory Service (MIDS) have been created to determine a network's latency, or time between nodes. Other studies and experiments such as Mapnet have attempted to map positions of the Internet. These studies and experiments however provide neither real analytical analysis capabilities nor practical application to the Internet infrastructure.
Wheeler and O'Kelly differ from previous approaches by researching the efficiency of the Internet. The study was divided into two categories--network connectivity and city accessibility. To study this, the two created a database of 31 Internet backbone providers that primarily serve the United States. Using these 31 Internet providers they applied three established formulas for network analysis--the gamma index, a T-matrix, and a D-matrix. According to the article, the gamma matrix is used to calculate the overall connectivity of the provider. The T-matrix is the measure of the total number of direct and indirect connections to a node network. The D-matrix is a measure of the shortest path between nodes, or points of origin and destination.
The data of the 31 providers was applied to a matrix format named the C-matrix. This matrix indicates the presence or absence of direct linkage between nodes. The C-matrix is then multiplied by itself over again until it reaches the diameter of the network being studied. The final matrix represents the T-matrix.
Taking the information from the T-matrix the two were able to determine the D-matrix. For every C-matrix multiplied in the T-matrix formula, a D-matrix...