Over the last few years graph databases have been the fastest growing sector within the database market (see db-engines.com/en/ranking_categories) and RDF databases, which are listed separately but are actually a subset of the overall graph market, have also been growing significantly. Unfortunately the site’s listings are not complete, and neither is it wholly accurate. The same applies to the comparable Wikipedia site.
Specifically, graph analytics and graph visualisation are topics that are attracting increasing attention both from IT vendors and from users. Why? What is it about graphs that makes them useful? For what sorts of application and use cases might it be sensible to use graphs? Couldn’t you use conventional technology to achieve the same thing? Indeed, what are graphs anyway?
This paper answers all of these questions. However, it isn’t just a matter of deciding “oh yes, let’s do graphs” because there are different approaches to the implementation of graphs and these have important implications for performance and scale that will impact on what you can and can’t do with particular technologies. So we also discuss the architectural considerations – at a relatively high level, this is not intended to be a paper for technologists – of how you implement graph technology.