Graph evolution from the 90s
I have vivid memories of professors using words such as vertexes and nodes during my first courses in discrete mathematics at university level. I left the courses feeling puzzled as I had never thought before drawing circles and lines between would be so complex. It turned out “travelling salesman problem” and many other interesting algorithmic problems all operate on the same data model – that of graphs.
There are some variants whether the graph nodes have directed connections or not and whether the connections between the nodes of the graphs can have properties of their own. These variants are all available in software based implementations where some real-world domains are better supported than some others and some algorithms that would optimize business (like that of the travelling salesman) are more efficiently evaluated than some others.
Since my university days popular Web sites such as Linkedin and Facebook have brought the notion of a graph and connectivity to the masses. People are connected to former and current colleagues as well as with companies. Friends are connected to not just friends but with their favourite sports or holiday resorts as well. Naturally companies are trying to extract the business value from these connections to their users’ benefit.
I have seen approaches such as deductive databases where the graph can be modelled as a logic program using binary relationships, I’ve seen topic map implementations, I’ve seen proprietary implementations and I’ve guided myself a semantic web based implementation at Profium. What is common to all of them is the ability to draw on the white board something which is easily understood by human beings and then mapped to a formal representation for processing with software.
Say a company is governed by a board which is composed of board members. The board members are further connected with other companies. A decision made at one of the companies can then be considered to be known (NDA’s to be respected of course) at other connected companies via their board members. A business intelligence solution can, for example, help a user to better understand if a fraud at one company could be suspected to happen at connected companies.
When your business decides to harness the value of connections, what remains for you to understand is the type of graph database you need. I would personally be happy to help you evaluate Profium Sense graph database suitability for your business data management challenges.