The main feature of scale-free network is the presence of hubs, that are nodes with two peculiarities:
- The number of hubs is much smaller than other nodes one, according to a power-law distribution.
- The number of links per hub is much greater than normal nodes one.
The Internet is the clearest example of such a network, however in my opinion there’s a big difference between this model (which is much more accurate than Erdos’ one) and several real cases of scale-free networks. On the Internet there are lots of hubs, but the majority of them is made up of search engines like Google, Yahoo, MSN, Alexa, etc. which don’t have any real link to other resources !
First of all, the number of crawled pages can be considered as a parameter of a function to obtain the (real) number of outcoming connections, because a concrete link is established only when a user select a website in order to visit it. Instaed theorical (and simplest) model of scale-free networks assures that every destination node can be reached with the same effort and probability; of course this condition is true for a large airport where any flight can be accessed by adopting the same procedure, but everybody knows that even a site has been crawled, only a limited number of keywords (for example, its domain name) are suitable for the appearance of a specific result.
This condition can be modeled with the introduction of “fluid” or “real-time” hubs: they are nodes which are able to meet any user’s requirements (the possibility to reach another element) using their specific features or skills. They could not be really connected with any possibile destination, however their abilities are enough powerful to allow particular real-time requests (in other contexts the need of time may be quite different, however it’s always less than a standard request fulfillment needs) giving the impression to be real enormous hubs.
Many companies have adopted this kind of strategy in order to increase their marketing efficiency: even they cannot satisfy any possible request, they say to be able to do that… In other words they become fluid hubs because the context where they operate is made up of small nodes which needs are seldom concurrent. Only a limited group of users will ask for a specific service, while the majority of them will be satisfied with “common” and easy to get solution.
In this way there’s no need to be real hubs (maybe it could be impossible) because a sufficient investment (both economical and technical) can assure the possibility to appear greater, stronger and faster than other “standard” competitor.
Share This Sphere: Related Content
hubs