The online social network poses severe challenges because of its large size and complex structure. It is meaningful to construct a concise high-level architecture of the online social network. The concise high-level architecture was composed of the communities, the hub nodes and the relationships between them. The original online social network was represented by a new representation named quantitative attribute graph, and a new method was proposed to construct the concise high-level architecture of the online social network. The communities were detected by using the attributes of the nodes and edges in combination, then the hub nodes were identified based on the found communities, and the relationships between the communities and hub nodes were reproduced. The new method was used to construct the concise high-level architecture of a large online social network extracted from a practical business Bulletin Board System (BBS). The experimental results show that the proposed method has a good performance when the relationship strength and the community size are set as 0.5 and 3 respectively.