The Internet is growing by Moore's law

The Laws of Mur is an empirical observation made in 1965, in the process of preparing the speech by Gordon Moore (one of the founders of Intel). He suggested that the number of transistors on a chip will double every 24 months. Having represented in the form of a graph, an increase in the performance of storage microcircuits, he discovered a pattern: new models of microcircuits were developed after more or less identical periods (18-24 months) after the appearance of their predecessors, and their capacity increased approximately twice each time. If this trend continues, Moore concluded, then the power of computing devices will increase exponentially over a relatively short period of time.

Today, a team of Chinese scientists has proven that the Internet is also growing according to Moore's law. In a recently published study, scientists concluded that the global network doubles every 5.32 years.


This is one of several discoveries published by researcher Guo-Qing Zhang and colleagues in the latest issue of the New Journal of Physics. Scientists have carefully analyzed the evolution of the global network topology, especially its structure and connections. In the course of the study, conducted from 2001 to 2006, scientists not only predicted how exactly the Network would expand - they were able to predict in which direction it would evolve.
')
Scientists viewed the Internet not from the point of view of WWW sites and their numerical increase and expansion, but from the point of view of autonomous systems (AS). As the project manager G. Jung explained, autonomous systems are self-sufficient regions that can exist without the influence of external factors on them. A university network, an Internet provider or a large corporate network can all be defined as AS.

In the model proposed by the researchers, each node (network node) is represented as an autonomous system. The connection between two independent NPPs is actually represented by both a physical connection (fiber, wires) and a commercial connection (contract, cooperation agreement).

To determine where the “core” of the Internet, which is its main component, is located, researchers have resorted to a method called “k-core decomposition”, that is, all nodes with a significance lower than the “K” coefficient have been removed from the virtual model Network. It turned out the kernel with a very small number of remaining nodes - with the central part (the so-called nucleus) less than 0.3% of the existing nodes are connected.

Scientists have found that the core of the Internet and its peripheral systems are subject to unequal evolutionary mechanisms. While the center of the Network is stable in its development, most of the new connections appear at a sufficient distance. This result is contrary to previous studies, which proved that the maximum number of connections with the “K” coefficient increases with the growth of the Internet as a whole.

“Many models have already been created that allegedly predict the growth of the global network in the future, and they all assume that the central part of the Internet is growing along with its periphery, depending on the same mechanisms. And all this is happening at a tremendous speed, ”- comments Jang -“ Our studies, however, proved that the maximum degree (number of connections) and the size of the central core of the Network remain the same - all growth occurs in the periphery. These results are fundamentally contrary to previous work on this topic, and they can be considered pivotal in modeling the global network. ”

Probably the most accurate model for describing the evolution of the Internet could be the so-called “positive feedback” (PIC) model (positive-feedback). According to this model, the number of communication nodes is increased by feedback cycles from node degrees. However, scientists admit the presence of yet unknown mechanisms by which the number of “root” “K” compounds remains static.

Studies have also proven that the Web is much more unevenly distributed than previously thought. Although the number of connections with large nodes is quite high, the results clearly showed that there are much more connections between small and medium nodes.

Scientists explain that understanding the topology of the Web is vital for optimizing its work. They hope that their research will bring practical benefits and help create the most complete model of the Internet.

“Understanding the network topology is of utmost importance when creating routing protocols, optimizing P2P traffic, and dealing with cascading collapses and overloads,” said study co-author Tao Zhou.

Via physorg.com