NETWORK EFFECT OF SEMANTIC WEB

NETWORK EFFECT OF SEMANTIC WEB

2011, Dec 24    

In brief, Semantic Web is a network of distributed databases as oppose to existing web which is more of a network of distributed webpages. According to Robert M Metcalfe (also known as Bob Metcalfe), the value of a network is proportional to the square of the number of nodes on the network. Mathematically it can be expressed as

V = n^2 (where, “V” is the value of the network and “n” is the number of nodes in that network)

Metcalfe’s law is also known as “Network Effect”. Even though initially Robert M Metcalfe formulated this law for Ethernet network however it has been implemented in other networking concepts as well that include web technology, social networking, business network and so on.

Can we can also apply Metcalfe’s network effect law on Semantic Web to estimate its value? Semantic Web is still just a vision, however we can consider Linked Open Data projects as miniature versions of Semantic Web. Therefore, by estimating the network effect of Open Data we can get an idea about the value of Semantic Web. By definition, Linked Open Data are obviously connected therefore we can consider Linked Open Data as a network of Open Data.

Now, as most of the Linked Open Data projects follow W3C Semantic Web recommendations and use RDF therefore, we can consider those individual data as “nodes” and the Semantic Web as a network. So, every time a new data or dataset added as Linked Open Data the value of Semantic Web is increasing according to Metcalfe’s law. Below is a visualisation of open Linked Data sources (we can consider them as “nodes” in Metcalfe’s law) of DBPedia.org –

Image

Figure: Some of the data sources of DBPedia.org

In Web 2.0 literature, Metcalfe’s law is often used to emphasise the value of social networks. In Semantic Web, the value of the Linked Data network would be a lot more than web 2.0 (or social web) as in this case number of nodes (data or dataset) is much bigger.

If we take Reed’s law into consideration; for defining the value of Semantic Web we will get even a larger valuation than that we get using Metcalfe’s law. In Reed’s law, value of a network grows much faster than Metcalfe’s law. According to Reed’s law, the value and power of a network increases exponentially with the number of nodes in the network; expressed mathematically as 2 to the nth power (2n). That means –

V = 2n (where, “V” is the value of the network and “n” is the number of nodes in that network)

In Reed’s law, the number of possible sub-groups within a network also being taken into consideration. Again, social networking websites can be an ideal analogy to explain this growth. In social networks (e.g. Facebook) we can form “groups”. In Metcalfe’s law these the number of “groups” is not considered, only the number of “members” is considered. However, obviously the members can also form group within the social network and these groups can also add value to the network in much larger scale than the value created by a single individual member.

Mashup applications is a notable use of Open Linked Data projects. In Semantic Web, we can take the different datasets from different sources and develop more mashups applications. Every time a new Linked Open Dataset is added in Semantic Web, it also adds a possibility of forming a new group with other dataset(s). Therefore, with inclusion of a new Linked Open Dataset in Semantic Web, the value of Semantic Web increases in a huge scale (2n) if we apply Reed’s law.

However, Reed’s law is not also out of any criticism. Obviously, all the Linked Data are not relevant to each other (even though technically they can be linked). Therefore, critics say Metcalfe’s law and Reed’s law overestimate the value of networks to some extent.

In one of my previous posts, I sort of divided all the Open Data projects into two categories – Linked Open Data projects (e.g. DBPedia, Data.gov.uk etc) and Non Linked Open Data projects (Pachube, DataGM etc). So according to our above discussion by applying Metcalfe’s law and Reed’s law we can say DBPedia and Data.gov.uk have more network value than Pachube and DataGM as the Open Data of first two websites are Linked Data. Now if we can transform the entire web into Semantic Web then the network effect of the web will be astronomical which will enable enormous number of new applications of the web.


Image source: Lifeboat.com


Love traveling? Visit my travel picture blog Tales of Trails