Multipoint communications is the simultaneous transmission of data streams from a number of sources to a set of receivers in a group according to predetermined metrics. The core-based approach in multipoint communication enhances potential solutions in terms of quality-of-service (QoS)-efficiency and feasibility of the results in inter and intra-domain routing. In this paper, we first analyse the solution space for constrained multipoint communication problems under the core-based approach. We show that the range of solutions examined by the models proposed to date is restricted to a subset of the entire solution space, which restricts the potential efficiency of the results. We propose SPAN, a core-based framework processing on our identified extended solution space for constrained multi-source group applications. SPAN consists of core selection and tree construction as two modular components complimenting one another to achieve more efficient solutions in distributed processing. SPAN is also asymmetric, hence potentially operates in domains in which link weights are not necessarily equal in both directions. We analyse the computational and message complexity of our framework and show its feasibility for distributed deployment. Our performance results show that SPAN consistently outperforms its counterparts in the literature in terms of cost and QoS-efficiency. Copyright (c) 2007 John Wiley & Sons, Ltd.