Optimizing wireless network throughput via network coding and placement of small cells

Xiaohang Li, Purdue University

Abstract

The growing popularity of mobile data services have placed higher and higher traffic demands for wireless network. Today's multimedia streaming applications not only require high throughput, but also stringent Quality-of-Service (QoS) guarantees. One way to support high-throughput applications is to design better scheduling algorithms for wireless network. We investigate the use of network coding along with the scheduling algorithms for delay-sensitive multimedia applications. Although network coding has been shown to significantly improve the throughput of communication networks, the existing design of network coding algorithms often does not account for delay, and hence may in fact lead to poor delay performance. How to design network coding algorithms with delay constraints thus becomes a challenging problem. We focus on a single-hop wireless setting with packet erasures. We first study streaming broadcast of stored-video over the downlink of a single cell. We generalize the existing class of immediately-decodable network coding (IDNC) schemes to take into account the deadline constraints. We also provide the deadline-constrained capacity-achieving IDNC scheme for a two-unicast scenario. Note that in both the broadcast scenario and the two-unicast scenario, the performance analysis of IDNC schemes is significantly complicated by the packet deadline constraints (from the application layer) and the immediate-decodability requirement (from the network layer). Another way to meet the rapidly growing traffic demand is through heterogeneous network deployment, which uses a mixture of macro cells and small cells (i.e., micro-/ pico-cells) to further enhance the spatial reuse and improve network throughput. We propose a Gibbs-sampling based optimization method for finding the optimal deployment of a given number of small cells in 3G networks. The Gibbs Sampling method intelligently balances two potentially conflicting considerations of placing small-cell base stations (BSs) close to hotspots and avoiding interference with the macro-cell BSs & other small-cell BSs. We show that it converges to the deployment decision with the maximum total system throughput with high probability. Our numerical results show that the Gibbs sampling-based placement outperforms the greedy solutions.

Degree

Ph.D.

Advisors

Lin, Purdue University.

Subject Area

Electrical engineering

COinS