Efficient sharing of radio spectrum has been a central focus of the wireless research community for some years [9], [11], [28], [34]. Most research on spectrum sharing has followed the socalled interweave paradigm [10], under which the secondary users’ activities do not overlap with the primary users’ in time, frequency, or spatial domains. Another implicit assumption under the interweave paradigm is that there is minimal cooperation between the primary and secondary networks on both the data and control planes. Recently, there is a growing interest in exploring cooperation between primary and secondary networks. For example, in [2], [10], [13], [14], [20], [21], [25], [26], [29], [33], the authors explored the benefits of unilateral cooperation, i.e., to have secondary users help relay traffic for the primary users. To take cooperation one step further, in [31], [32], the authors advocated a bilateral cooperation between primary and secondary networks, where the two networks can help relay each other’s traffic. Such bilateral cooperation allows to pool together the resources from both networks so that users in each network can access a much richer network resources from the combined network. It was shown in [31], [32] that such bilateral cooperation brings many potential benefits and flexibilities on both the data and control planes that are otherwise not possible. Note that although the two networks are combined into one at the node level, priority or service guarantee to the primary network traffic can still be enforced by implementing appropriate traffic engineering rules.