Packet-Level Modeling of Cooperative Diversity: A Queueing Network Approach

Cooperative communications are expected to be a centerpiece for 5G cellular networks. Using cooperation, a wider and more uniform network coverage is attained, network capacity is enhanced, and power consumption is drastically reduced. While these latter heavily rely on collaboration among network elements, there is barely a packet-level perspective on what tangible gains cooperation is able to achieve. Motivated by this fact, different from studies on the capacity analysis of cooperative protocols, this paper tackles cooperation as a packet-level problem. The latter perspective empowers us to broaden our understanding of cooperation through characterizing high-level quantities, such as delay, throughput, fairness, and buffer length, as more tangible measures of the instant service quality that users/devices experience. In the pursuit of achieving this goal, the theory of BCMP queueing networks is leveraged. The proposed modeling approach can be used to analyze networks with an arbitrary number of relays, traffic classes, generic service time distributions, and several serving disciplines. To showcase the generality of this framework, we establish the queueing models for some of the most well-known cooperative protocols, such as amplify and forward amplify-forward, decode-forward, selection-relaying, incremental-relaying, and opportunistic-relaying, and evaluate their performances through the above-mentioned metrics. Moreover, a distributed cooperative protocol based on space-time block codes is proposed, its corresponding BCMP model is derived, and its performance is compared with other cooperative protocols.