Coordinated Multi-point Transmission & Reception in Heterogenous Networks
In this work we consider Coordinated Multi-Point transmission and reception (CoMP) schemes over heterogeneous wireless networks. These heterogeneous networks comprise of a set of disparate transmission points serving multiple users on an available spectrum. To enable better resource allocation, the set of transmission points is partitioned into several clusters and each cluster is assigned a set of users that it should serve. Joint resource allocation (scheduling) using all transmission points in a cluster and a suitable CoMP scheme is possible due to the availability of fiber backhaul within each cluster. Our contributions are in the design of approximation algorithms for this joint scheduling problem. We show that the joint scheduling problem is strongly NP-hard and then design an approximation algorithm that yields a constant factor approximation. To further obtain algorithms with a substantially reduced complexity, we adopt an iterative framework and design three polynomial time approximation algorithms, all of which yield constant factor approximations for a fixed cluster size. The design of these algorithms also reveals a useful connection between the combinatorial auction problem with fractionally sub-additive valuations and the submodular set-function maximization problem. We then conduct a thorough evaluation using models and topologies developed by the 3GPP standards body to emulate such networks. Our evaluations show that by exploiting all the feedback provisioned in the standard in a certain manner and by using well-designed algorithms, significant CoMP gains can be realized over realistic heterogeneous networks.
NemoX: Scalable Network MIMO for Wireless Networks
Network MIMO (netMIMO) has potential for significantly enhancing the capacity of wireless networks with tight coordination of access points (APs) to serve multiple users concurrently. Existing schemes realize netMIMO by integrating distributed APs into one “giant” MIMO but do not scale well owing to their global synchronization requirement and overhead in sharing data between APs. To remedy this limitation, we propose a novel system, NEMOx, that realizes netMIMO downlink transmission for large-scale wireless networks. NEMOx organizes a network into practical-size clusters, each containing multiple distributed APs (dAPs) that opportunistically synchronize with each other for netMIMO downlink transmission. Inter-cluster interference is managed with a decentralized channel-access algorithm, which is designed to balance between the dAPs’ cooperation gain and spatial reuse—a unique tradeoff in netMIMO. Within each cluster, NEMOx optimizes the power budgeting among dAPs and the set of users to serve, ensuring fairness and effective cancellation of cross-talk interference. We have implemented and evaluated a prototype of NEMOx in a software radio testbed, demonstrating its throughput scalability and multiple folds of performance gain over current wireless LAN architecture and alternative netMIMO schemes.
NEMOx: Scalable Network MIMO for Wireless Networks
The 19th Annual International Conference on Mobile Computing and Networking (MobiCom 2013)
pp. 453-464, 2013
Xinyu Zhang, Karthikeyan Sundaresan, Mohammad A. Khojastepour, Sampath Rangarajan, Kang G. Shin