DHHTGraphs - Modeling beyond plain graphs

Graphs are known as a suitable representation of structured data in general and of models in particular. Most common graph databases and storage facilities are based on a relatively simple kind of graphs restricted to binary and often even untyped or non-attributed edges. However, for domain-specific modeling there is also a need for some natural advanced concepts such as n-ary edges, several kinds of hierarchy and distribution of a graph over several sites. While common graph approaches require to simulate those concepts by plain graphs, this paper introduces an enhanced approach offering a seamless integration of these concepts, called Distributed Hierarchical Hyper-TGraphs (DHHTGraphs). Based on a formal definition, DHHTGraphs are described in detail including a metamodeling language and an efficient implementation.