ABSTRACT
In this chapter, the authors introduce the basic concept of big graph, a special type of big data. They propose an efficient scheme to decompose a large graph into subgraphs, and suggest a Merkel tree-based hashing scheme. Big graph, also known as large-scale graph-structured database, is a popular solution for representing social networks, healthcare records, and Internet of Things data. There are mainly four types of graphs, that is, the tree, the directed acyclic graph, the graph with cycles, and the graph with multiple source vertexes. The basic philosophy behind think like a vertex is to iteratively execute the graph algorithms over vertexes, which means that a vertex program only handles data from its adjacent vertices and incident edges. In graph-based machine learning algorithms, the data in the graph is frequently accessed and updated. Thus the most important security requirements in a graph-structured database are access management and integrity tests.
