1 Introduction
Graph data are key parts of big data and widely used for modelling complex structured data with a broad spectrum of applications. For example, the explosion of social network in the last decade leads to very large scale graph data. In this paper, we focus on the triangle counting in graph data [1], [2]. Essentially, in a graph G, a triangle is a set of three nodes among which each pair of nodes are neighbours in G. Ever since Watts et al. [3] introduced the measurement to determine whether a graph is a small-world network, many work focus on finding the triangles, which are one of the simplest but effective descriptions of a node’s status in a small-world [4], [5], [6].